close

Вход

Забыли?

вход по аккаунту

?

Time domain dielectric microwave detection of biomolecular surface interactions with a coplanar transmission line probe

код для вставкиСкачать
Time Domain Dielectric Microwave Detection of Biomolecular
Surface Interactions with a Coplanar Transmission Line Probe
By
Qin Chen
B.S. (East China University o f Science & Technology) 1997
M.S. (University o f California, Davis) 2003
DISSERTATION
Submitted in partial satisfaction of the requirements for the degree of
DOCTOR OF PHILOSOPHY
in
Electrical and Computer Engineering
in the
OFFICE OF GRADUATE STUDIES
of the
UNIVERSITY OF CALIFORNIA
DAVIS
APPROVED:
Committee in Charge
2005
i
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3191104
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3191104
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Copyright by©
QIN CHEN
2005
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
If life is like a journey, people always would look back at certain point during the
journey. This thesis is a result o f four and a half years work. When I look back, I find that I
have been accompanied with many wonderful people. So, it is a great pleasure to have an
opportunity here to express my gratefulness.
First, I am deeply indebted to my adviser Prof. Andre Knoesen. His constant
encouragement, stimulating suggestions and constructive critique accompanied with me
throughout the research and writing o f the thesis. Furthermore, he showed me his great
persistence in pursuing excellent work, and strong personality, which is very important in
research and in life. He may not realize how much influence he made on me, but I am very
glad that I met Prof. Knoesen in my life.
I would like to thank my mentor Dr. Daniel Roitman from Agilent Laboratories,
who has been spending amazing amount of time and effort to guide me through the research
and always been supportive. It has been an honor and great pleasure to work with a brilliant
scientist like him. I would like to thank Prof. Heritage for reading the thesis and giving me
constructive feedback, and also for every little chat, which either gave me encouragement or
made me appreciate tackling difficult problems with easy fundamental approaches. I have
been extremely fortunate to work with these great research advisers, whose enthusiasm in
science inspired me.
I would like to thank all my group members who have helped me along this journey,
including: Prof. Diego Yankelevich, Dr. John Campbell, Dr. Carl Arft, Dr. Israel Rocha and
Mingshi Wang. Prof. Diego Yankelevich who is an excellent researcher and teacher,
provided me help and warm encouragement all the time, together with good chocolates and
funny jokes. Dr. John Campbell helped me set up the first testing system, and kept on
providing me support even after he left Davis. Dr. Carl Arft developed the FDFD modeling
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
tool in his thesis work, which was used in modeling the CTL probe. I would also like to
thank him for fixing my car and my bike more than once.
This research project is a collaboration between UC Davis, Agilent Technologies,
and IBM Almaden Research Center. I have been fortunate to have the opportunity to interact
with excellent researchers from different fields. I would like to thank those who made my
research work possible. Maggie Bynum, Dr. Daniel Sobek, Dr. Kelvin Kleen and Jean
Norman (Agilent Laboratories, Palo Alto) helped me with the bio-experiments. Hassan
Tanbakuchi, Roger Stancliff (Agilent Technologies, Santa Rosa) helped me with microwave
measurement equipments. Dr. Robert Miller, Dr. Ho-Cheol Kim, Sarah Angelos (IBM
Almaden Rsearch Center) provided nanoporous film deposition. Prof. Gangyu Liu and Dr.
Maozi Liu (Department o f Chemistry, UC Davis) helped me with the AFM measurements. I
had the pleasure to supervise and work with several undergraduate students who did summer
research in our group and helped me with this research project, including: Daniel Clark,
Michael Lei, Alexander McCourt and Joey McMurdie.
I would like to thank a group o f friends, who I got the chance to know during my
stay in Davis. Sarah Teter, Jennifer Epp, Chris Lupo, Mark Heffeman, Rajinder Sahota,
Honey Walters, Steve Zelinka, Anne Donnelly, Brian Donnelly, Jennifer Fleck, Matt Fleck,
Steve Pigg, Jing Zhang, Keyao Zhu, Hui Zang, Jun Xiu, Yongfang Guo, Hong Wong,
Cameron Blatter, and Jessi Johnson. They brightened my journey like flowers.
Last but not least, I would like to thank my family. My parents who teach me good
things that really matter in life, never lose confidence in me and always care about me. My
sister and my brother-in-law are always proud o f me and encourage me in difficult times.
Without their support, I could never make it this far. This dissertation is dedicated to them.
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract
Developing new techniques to detect biomolecular interactions (BMIs) taking place
at a liquid/solid interface is o f great interest for drug discovery, clinical diagnostics,
and environmental analysis. This dissertation project covers the development o f a
new technique that uses a microwave coplanar transmission line (CTL) to detect the
biomolecular interactions occurring at the solid/liquid interface. This new technique
is label free, non-destructive, and could detect the biomolecular interactions in the
native environment.
The protein ligand molecules are chemically immobilized on the glass region o f CTL
surface. The transmission line is exposed to a saline liquid solution containing a
variety o f target biomolecules. Only specific target molecules in the fluid sample
bind with the ligand molecules to form the protein conjugate complex at the
transmission line interface. The reaction results in a subtle change in the dielectric
constant at the solid/liquid interface. In this thesis, it is demonstrated that even
though this permittivity change is small, it can be detected by measuring the
dispersion property change o f the electromagnetic signal at the frequency range from
a few Hertz to around 25 GHz. Time domain measurement was made, and the pulse
transit time shift was used to quantify the effective dielectric constant change.
The CTL probe was used to investigate different BMIs, such as the biotin and anti­
biotin (anti-biotin mouse y-immunoglobulin), and protein A and antibody (rabbit yimmunoglobulin) interaction. The results showed that the probe was sensitive to the
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
dielectric property changes within a few molecular layers. The sensitivity o f the
probe reached xOO pg/mm . The technique could also perform a real-time
measurement o f the BMIs.
Different parts o f y-immunoglobulin were involved in these two BMI assays (biotin
to anti-biotin mouse IgG and PA to R-IgG). The CTL probe showed different
sensitivity to these two binding reactions. CTL probe coupled with other detection
techniques, fluorescence and atomic force microscopy, could gather more complete
information about a BMI, such as the protein orientation, protein package density,
and protein-water interactions.
The BMI event was simulated as a thin non-dispersive dielectric layer displacing
water at CTL the interface. The simulation was done with an electromagnetic model
using the FDFD-CP technique, and was verified by a deposition o f the
polyelectrolyte multilayer experiment. The study indicated the probe could detect a
dielectric layer as thin as 2nm displacing water on the transmission line.
The nanoporous thin film was applied to the CTL probe surface to improve the
loading of ligand molecules. The probe showed selectivity to the target molecules
(avidin, and anti-biotin mouse IgG) based on their physical sizes.
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS.......................................................................................... i
ABSTRACT.................................................................................................................. iii
TABLE OF FIGURES& TABLES.............................................................................v
1. INTRODUCTION.................................................................................................... 1
1.1 Background.............................................................................................................................. 1
1.2 Scope o f the Dissertation....................................................................................................... 3
1.3 References................................................................................................................................ 7
2. DETECTION OF BIOMOLECULAR INTERACTIONS.................................9
2.1 Anatomy o f Biomolecular Interaction Sensors.....................................................................9
2.2 Basic Permittivity Concepts.................................................................................................. 12
2.2.1 Polarization and Polarizability o f the Material.....................................................12
2.2.2 Dielectric Relaxation and Complex Permittivity..................................................13
2.3 Dielectric Property o f Biological Materials........................................................................ 14
2.4 Transmission Line Dielectric Measurement Techniques...................................................16
2.4.1 Coaxial Transmission Lines.................................................................................... 17
2.4.2 Planar Transmission Line as Dielectric BMI Probe............................................. 18
2.5 Planar Transmission Line Surface Biomolecular Deposition Schemes.......................... 20
2.6 Conclusion...............................................................................................................................21
2.7 References...............................................................................................................................23
3. COPLANAR TRANSMISSION LINE PROBE IMPLEMENTATION AND
DETECTION OF BIOMOLECULAR SURFACE INTERACTIONS.........25
3.1 Introduction............................................................................................................................25
3.2 Coplanar Transmission Line Probe.....................................................................................26
3.2.1 Coplanar Transmission Line/ Microfluidics Design and Fabrication.............. 26
3.2.2 Time Domain Transmission System Set-up.........................................................30
3.3 Feasibility Test Experiments................................................................................................31
3.3.1 Feasibility Test 1- Specific Binding Experiments: Biotin to Anti-biotin
and Rabbit IgG to Anti-rabbit Ig G .........................................................................31
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.3.1.1 Experimental........................................................................................... 32
3.3.1.2 Results..................................................................................................... 33
3.3.1.3 Discussions............................................................................................. 37
3.3.2
Feasibility Test 2- Kinetics Study and Affinity Constant Estimation o f Protein
A and Rabbit IgG.................................................................................................39
3.3.2.1 Introduction..............................................................................................39
3.3.2.2 Experimental........................................................................................... 40
3.3.2.3 Dielectric Measurement Results and Fluorescence Results............. 42
3.3.2.4 Sensitivity Estimation and Affinity Constant Calculation.................46
3.4 Conclusions..........................................................................................................................48
3.5 References............................................................................................................................50
4. DIELECTRIC COPLANAR TRANSMISSION LINE PROBE MODELING
AND SENSITIVITY STUDY................................................................................. 51
4.1 Introduction.........................................................................................................................51
4.2 Theoretical Model for CTL Probe....................................................................................52
4.3 Experimental Verification o f the M odel......................................................................... 61
4.4 Results and Discussions.................................................................................................... 63
4.5 Conclusion...........................................................................................................................66
4.6 References...........................................................................................................................67
5. CHARACTERIZATION OF BIOMOLECULAR INTERACTIONS USING
COPLANAR TRANSMISSION LINE PROBE COUPLED WITH
FLUORESCENCE TECHNIQUE......................................................................... 69
5.1 Introduction....................................................................................................................... 69
5.2 Experimental...................................................................................................................... 71
5.3 Results................................................................................................................................. 73
5.4 Discussions..........................................................................................................................79
5.5 Conclusions.........................................................................................................................84
5.6 References...........................................................................................................................86
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6. NANOPOROUS THIN FILM ON THECTL TRANSMISSION LINE
PROBE SURFACE FOR BMI DETECTIONS.................................................. 88
6.1 Introduction.........................................................................................................................88
6.2 Experimental...................................................................................................................... 89
6.3 Results................................................................................................................................. 93
6.4 Discussions..........................................................................................................................96
6.5 Conclusions........................................................................................................................ 98
6.6 References...........................................................................................................................99
7. CONCLUSIONS AND FUTURE WORK..........................................................100
7.1 Conclusions.......................................................................................................................100
7.2 Future Work.......................................................................................................................106
7.3 References.........................................................................................................................I l l
APPENDICES............................................................................................................ 113
A. Debye Model and Cole-Cole Model..............................................................................113
B . Maxwell-Wagner Interfacial Polarization..................................................................... 119
C. Calculation for Permittivity o f Saline Water................................................................123
D. Surface Imaging of Antigen-Antibody Interactions on Biotinylated CTL Probe
Using Atomic Force Microscopy................................................................................... 125
D .l Experimental............................................................................................................125
D.2 Results and Discussions...........................................................................................126
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF FIGURES & TABLES
Figure 2.1
Coaxial cells for dielectric measurement in microwave regime
Figure 2.2
Coplanar transmission line configuration...........................................20
Figure 3.1
Three-layer coplanar transmission line (CTL) configuration with (1)
glass substrate ( s r = 4 .6 ) (2) a thin aqueous sample layer
18
(replaced with air in design) and (3) the PMMA sample cover
( s r = 2 .7 ) .............................................................................................. 27
Figure 3.2
Frequency response (S 2 1 ) of the CTL probe with air on top,
measured with the Agilent 8703A Network Analyzer..................... 27
Figure 3.3
A cross section o f the CTL probe. Ligand protein molecules were
immobilized onto the glass interface between the signal and ground
electrode. Target molecules flow through the fluidic channel and
bind selectively to the ligands, which lead to a change in
permittivity............................................................................................ 29
Figure 3.4(a)
The pulse transmission through PBS buffer before (dashed line) and
after (solid line) the biotin-anti,biotin binding reaction...................34
Figure 3.4(b)
The pulse trace difference of PBS buffer base line (dashed line) and
the trace difference after the biotinylated surface exposed to anti­
biotin....................................................................................................... 35
Figure 3.4(c)
The pulse trace difference o f the biotinylated surface exposed to
anti-biotin (solid line) and avidin (dashed line)................................35
Figure 3.5(a)
The CTL surface with R-IgG exposed to anti-mouse IgG(nonspecific binding)...................................................................................36
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.5(b)
The CTL surface with R-IgG exposed to anti-rabbit IgG
Figure 3.6
The fast pulse transmitted through CTL probe with PBS buffer on
top. The signal launched with SMA connectors pressed down on the
CTL (doted line) vs. the signal launched with SMA connectors
soldered on the CTL (solid line)..........................................................39
Figure 3.7
Estimating the pulse shift (Atp ) from PA-R-IgG binding using
cross-correlation, (a )two real pulses taken during the reaction, (b)
cross-correlation result o f the two pulses..........................................42
Figure 3.8
CTL probe real-time response to the exposure of
BSA (500pg/ml)....................................................................................43
Figure 3.9
Real-time binding curve of R-IgG on a PA-AMS surface. The
probe was exposed to R-IgG or PBS alternatively. The vertical
marker lines on the binding curve indicate the valve switching
points....................................................................................................... 45
Figure 3.10
UV fluorescence profile from the secondary antibody Cy3-labelled
anti-R-IgG ‘sandwich assay’ on top o f PA and R-IgG..................... 46
Figure 3.11
(R-IgG concentration/time shift) vs. (R-IgG concentration) from
which the overall affinity constant o f PA and R-IgG
36
are determined....................................................................................... 48
Figure 4.1
CTL probe geometry for the electromagnetic modeling..................55
Figure 4.2
Real
and
imaginary
parts
of
the
effective
dielectric
constant s eff —y 2as a function o f frequency and the layer thickness
for er - 2.5 . The effective dielectric constant was calculated from
the FDFD model....................................................................................56
Figure 4.3
Normalized pulse phase velocity as a function o f dielectric constant
and thickness of the thin dielectric layer............................................ 57
x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.4
Comparison of the theoretical model and measured pulses through
CTL probe with water in the fluid chamber. The theoretical model is
from simulating a 13ps FWHM Gaussian pulse propagating through
the CTL with dispersion properties calculated from the FDFD
model...................................................................................................... 58
Figure 4.5
Change in the bandwidth as a function o f propagation
length...................................................................................................... 59
Figure 4.6
Pulse attenuation as a function of propagation length.......................59
Figure 4.7(a)
A t as a function of dielectric constant and thickness o f thin
dielectric layer....................................................................................... 60
Figure 4.7(b)
T as a function o f dielectric constant and thickness o f thin
dielectric layer........................................................................................ 61
Figure 4.8
The raw experimental data o f dielectric probe response to the
polyelectrolyte multilayer deposition. The CTL probe was exposed
to 0.25% PSS (anionic) in 0.1 M NaCl and 0.25% PAH (cationic) in
0.1M NaCl alternatively. PBS buffer wash was applied after each
exposure..................................................................................................64
Figure 4.9
The raw experimental data o f dielectric probe response to the
polyelectrolyte multilayer deposition. The CTL probe was exposed
to 0.5% PSS (anionic) in 0.03M NaCl and 0.5% PAH (cationic) in
0.03M NaCl alternatively. PBS buffer wash was applied after each
exposure..................................................................................................64
Figure4.10
Time shift A t = t - t pei for the successive PSS and PAH layers
(using the time shift after the first PEI layer deposition as the
reference) a function o f number o f layers. The straight line is the
linear regression result from the data.................................................. 65
Figure 5.1
Two bioassay binding interactions take place at different functional
parts of y-immunoglobulin (IgGs). The protein A (PA) to RabbitIgG (R-IgG) interaction involves the Fragment crystallizable (Fc)
portion o f IgG. The biotin to anti,biotin mouse IgG interaction
involves the Fragment antigen binding (Fab) portion
o f the IgG................................................................................................70
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.2
CTL response to bovine serum albumin (BSA). Block solution
(1 mg/ml) BSA was introduced to the CTL bio-reaction channel for
10 minutes followed by a PBS buffer wash. The reference channel
was filled with PBS buffer during the experiment. The CTL
transient response to BSA bulk solution was At = -1.5ps. The pulse
shift after PBS wash was At = -0.1 ps................................................. 74
Figure 5.3(a)
The CTL probe surface modified with Protein A ligand(substrate(a))
was exposed to Rabbit IgG (R-IgG) as the analyte. The CTL
response was continuously monitored during the flow o f R-IgG
(Sigma C2821).The pulse shift reading (in the plot) was taken after
the buffer wash and before the next sample..................................... 76
Figure 5.3(b)
Fluorescence profile from Cy3-labelled R-IgG (Sigma C2821) on
PA modified substrate (a).....................................................................76
Figure 5.4(a)
The CTL probe surface modified with Protein A ligand(substrate (b))
was exposed to Rabbit IgG (R-IgG) as the analyte. The CTL
response was continuously monitored during the flow o f R-IgG
(Sigma C2821).The pulse shift reading (in the plot) was taken after
the buffer wash and before the next sample.......................................77
Figure 5.4(b)
Fluorescence profile from the Cy3-labelled -R-IgG (Sigma C2821)
on top o f PA modified substrate (b).................................................. 77
Figure 5.5 (a)
The CTL probe surface modified with biotinylated bovine serum
albumin (B-BSA).The ligand was exposed to anti-biotin mouse yimmunoglobulin IgG (a-biotin) as the analyte. The CTL response
was continuously monitored during the flow of a-biotin (Sigma
5585) solution o f increasing concentrations. The pulse shift reading
(in the plot) was taken after the buffer wash and before the next
sample..................................................................................................... 78
Figure 5.5 (b)
The UV fluorescence profile from the Cy3 labeled secondary
antibody a-biotin mouse IgG on the B-BSA surface was obtained
using Axon Instruments GenePix 4000A. The two feature peaks
with high fluorescence intensity confirmed that biotin binding was
predominantly inside the glass gap regions on the CTL. The
substrate yielded fluorescence o f 28,000 counts (PMT voltage =
500V)...................................................................................................... 78
Table 5.1
Summary o f Fluorescence and CTL Probe Results.......................... 79
X ll
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 6.1
Schematic o f the formation o f PMSSQ nanoporous film and SEM
image of the porous film.....................................................................90
Figure 6.2
The biotinylated CTL plain glass substrate response to a-biotin
mouse IgG (Sigma 5585). The CTL probe responded to a-biotin
IgG with increasing concentrations (500ng/ml, lpg/ml, 2.5pg/ml,
5pg/ml and lOpg/ml).......................................................................... 94
Figure 6.3
The biotinylated CTL plain glass substrate response to extra-avidin
(Sigma E4142) after exposed to a-biotin IgG (Figure 6.1). A total
time shift o f -1.53ps was observed after introducing the extra-avidin
with increasing concentrations (lOOng/ml, lpg/ml, and 5pg/ml) to
the substrate.......................................................................................... 94
Figure 6.4
The biotinylated nanoporous CTL substrate response to a-biotin
mouse IgG (Sigma 5585). The CTL probe had almost no response
to a-biotin IgG with increasing concentrations (500ng/ml, lpg/ml,
and 10 pg/ml), and very small pulse time shift was observed as
compared with the CTL probe without porous film (Refer to Figure
6.2)........................................................................................................ 95
Figure 6.5
The biotinylated nanoporous CTL substrate response to extra-avidin
(Sigma E4142) after exposed to a-biotin IgG. A total time shift o f 1.47ps was observed after introducing the extra-avidin with
increasing concentrations (lOOng/ml, lpg/ml, and 5pg/ml) to the
substrate..................................................................................................95
Figure A .l
Cole-Cole diagram displaying a semi-circle for Debye equations for
e*............................................................................................................115
Figure A.2
Equivalent Circuit for Debye Model and Cole-Cole Model
Figure B .l
Maxwell-Wagner model as two capacitors in series and equivalent
circuit....................................................................................................119
Figure D .l
The AFM scan image o f BSA-biotin surface reacted with
extravidin-Cy3 conjugate (half-wet). Notice the darker square area
at the right side corresponds to the AFM tip scratching mark on the
xiii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
protein surface. The height of the protein layer is
~6nm......................................................................................................127
Figure D.2
The line-plots o f the traces denoted in the image o f BSA-biotin and
extravidin. The height o f the BSA-biotin and avidin pair is
~6nm..................................................................................................... 127
Figure D.3
The AFM scan image o f BSA-biotin surface reacted with
extravidin-Cy3 conjugate (dried)....................................................... 128
Figure D.4
The line-plots o f the traces denoted in the image o f BSA-biotin and
extravidin. The height o f the BSA-biotin and avidin pair is ~6nm
which is the same as the result from the wet surface.......................128
Figure D.5
Another scan image of the BSA-biotin and extravidin surface.
Notice the o-ring features on the surface. The outer diameter o f ring
is ~90nm. Notice the ring shaped protein cluster is smaller for
avidin since an avidin molecule is smaller than an
antibody............................................................................................... 129
Figure D.6
AFM height image scan for BSA-biotin surface reacted with anti­
biotin mouse IgG -Cy3 conjugate................................................... 130
Figure D.7
The line-plots o f the traces denoted in the image o f BSA-biotin and
anti-biotin. The height o f the BSA-biotin and anti-biotin pair is
~17.5nm................................................................................................ 130
Figure D.8
A zoom-in AFM scan image for the BSA-biotin and anti-biotin
surface. Notice the o-ring shaped feature on the surface. The outer
diameter of the ring was measured to be ~150nm........................... 127
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Chapter 1 Introduction
1.1 Background
Biosensors that detect biomolecular interactions (BMI) are widely used to
map biochemical pathways that lead to disease states, monitor a patient for clinically
relevant analytes, detect infectious agents and environmental toxins, and develop
new pharmaceutical compounds. A BMI takes place when a molecule adheres to
another molecule, for example, enzymes to substrates, antibodies to antigens, or
DNA strands to their complementary strands. The binding occurs because the shape
and chemical properties o f the molecules are complementary. The Tock-and-key’ is
a common metaphor to describe how target molecules fit into their receptors.
To fully characterize and understand BMIs, knowledge about more than one
physical and chemical property o f the reaction is required. A BMI is typically
detected by measuring changes in only one physical or chemical property, such as
optical, viscoelastic or electronic properties. For this reason, developing new
biosensor techniques is always o f great interest for scientists and engineers.
Current biosensing techniques can be divided into two categories: molecular
labeling (also called tagging) techniques and label-free techniques. Molecular
labeling uses molecular tags (fluorescence, or radioactive labels) conjugated to one
or several interacting molecules. When a binding event takes place, the molecular
tagging signals are either induced or quenched. The fluorescence tagging method, a
common labeling technique, is one o f the most sensitive detection methods currently
available. A limit of detection o f femtomolar concentration has been reported [1].
However, the assays suffer from limitations imposed by photobleaching [2].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
Molecular tagging has some drawbacks. First, the chemical modification may
induce undesirable changes in the assay. For example, fluorescence labels are usually
hydrophobic [2] which may cause changes at the binding site o f the labeled
molecules. Second, a washing process is always needed before the measurement in
order to rinse away the excess labeled reagents, and minimize the background signal
from non-specific bindings. However, some valuable information such as the on/off
rate about the assay adsorption is lost during the washing process. Some in-situ
fluorescent technique exists, but requires a complicated optical set-up. Third, the
number o f molecular tags changes with the protein assay. For instance, each protein
type takes on a different number o f fluorephores depending on the number o f lysine
groups available in the protein. Thus, it is very difficult to make a quantitative
analysis [3].
Label-free detection methods detect the interaction through the perturbation
o f an electromagnetic or an acoustic mode. They address the limitations o f the
labeling techniques. A common optical label-free detection technique is surface
plasmon resonance (SPR). A SPR is a surface electromagnetic mode in which the
incident light on a metal-dielectric interface couples to oscillations o f electrons
within a metal film [4, 5]. For a SPR sensor, the sensing mechanism relies on
measuring the change in refractive index o f the surface as the analyte is introduced.
A SPR sensor is used to study the binding kinetics in real time, which is an important
attribute o f BMIs [6, 7].
Another label-free technique is the quartz crystal microbalance (QCM). The
QCM is a piezoelectric device fabricated on a quartz oscillator that is capable of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
measuring the mass change in nanogram range. The molecules adsorbed on the
quartz oscillator cause changes in mass loading on the sensor which changes the
resonant frequency of the oscillator.
One important aspect o f both SPR and QCM is the generic nature o f the
detection principle, i.e. they can not distinguish between different molecular
identities. The SPR response is derived from changes o f the refractive index at the
sensor surface. QCM responds to the BMI through detecting the total organic mass
change on the sensor surface. Therefore, special attention must be paid in designing
the binding experiments for specificity. A combination of label-free sensing with
complementary detection techniques must be used to obtain molecular structural
information. For instance, Biacore [8] is developing a sensor using SPR combined
with mass spectrometry to provide molecular identification along with binding
information [7, 9].
1.2 Scope of the Dissertation
This research project focuses on developing a new label-free biosensing
technique to detect BMIs at the solid-liquid interface. It is based on a microwave
transmission line structure to detect changes in electric permittivity near the
solid/liquid interface. It is a label-free technique, and can probe a biomolecule
directly in its native environment such as saline buffer. Research has shown that
biomolecules have distinctive dielectric features in the microwave range [10-17].
This additional molecular information not available in SPR or QCM suggests that
measuring the permittivity in the microwave frequency range should be superior.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
The organization o f this dissertation is as follows. Chapter 2 explains the
significance of the research and describes background knowledge pertaining to the
development o f the dielectric transmission line biosensor system. The SPR technique
is reviewed. Dielectric spectroscopy can be viewed as an extension o f SPR. Whereas
SPR measures changes in permittivity at optical frequencies, dielectric spectroscopy
measures changes at a lower frequency range in which the biological materials give
more spectral signatures. The factors that influence the dielectric constant in a
material are covered and the dielectric spectra o f biological solutions are summarized
to show what type o f molecular structural information can be obtained. Transmission
line dielectric measurement techniques are reviewed. The coplanar transmission line
is discussed to demonstrate that the configuration is suitable for fluidic handling, and
performing surface biomodification. The well-developed protein deposition
chemistry on semiconductor substrates is briefly discussed. We will focus on the
signal transduction technique development and incorporate the available surface
modification techniques and fluidic handling schemes to develop a new label-free
dielectric biosensor system.
Chapter 3 focuses on the coplanar transmission line (CTL) BMI probe
implementation and feasibility studies. We answer an important question: ‘Is it
possible to detect changes from BMIs within several molecular layers electrically
through measuring the permittivity change below 20GHz?’ Details o f the
configuration o f a CTL probe are presented. The surface modification with ligand
biomolecules on the CTL probe surface for detection o f specific BMIs is described.
Results of two sets o f feasibility measurements are presented to demonstrate the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
probe operation. The measurements were performed by immobilizing different
ligand molecules on the CTL surface and detecting the target molecules flow through
a liquid channel. The interfacial permittivity changes caused by the BMIs are
quantified by measuring the change o f the propagation time o f a fast pulse through
the transmission line. A real time measurement of a BMI is performed, and the
sensitivity o f the system is estimated.
Chapter 4 presents the CTL probe electromagnetic modeling. We
demonstrate that the sensitivity of the probe is due to an effective displacement o f
water molecules close to the liquid/solid interface. We measure permittivity changes
that occur at the liquid-solid interface by sensing changes in the dispersion o f the
CTL. A theoretical analysis of the dispersion changes is presented by modeling the
propagation o f a baseband Gaussian pulse. Experimental results are given that verify
the theoretical predictions in a CTL probe. The dielectric properties are
systematically changed at the liquid-glass interface by depositing oppositely charged
polyelectrolytes to form monolayers on the probe surface. We show that the
measured changes in pulse correlate well with the theoretical modeling.
Chapter 5 focuses on the application o f the CTL probe as a label-free
immunosensor in a BMI study coupled with the fluorescence detection technique.
We present results using two model assays. The first is protein A (PA) and anti-goat
rabbit y- immunoglobulin IgG (R-IgG), and the second is biotinylated bovine serum
albumin (B-BSA) and anti-biotin mouse y- immunoglobulin IgG (a-biotin) binding.
These two binding assays are chosen because the protein-ligand interactions involve
different parts of the y- immunoglobulin molecule, the Fragment crystallizable (Fc)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
portion o f IgG, and the Fragment antigen binding (Fab) portion o f the IgG. The
results illustrates that the CTL probe provides different response to the two different
binding assays. The different response suggests that the CTL probe can not only
detect the binding events but also is able to differentiate the binding events.
We also explore the potential o f using nanoporous structures to achieve better
sensitivity and selectivity. The result from biotin binding with two target molecules
with different sizes (anti-biotin and extra-avidin) suggests a more interesting
possibility to use a well-controlled nanoporous support with sufficiently uniform
nano pores to physically separate two target molecules based on their different
physical sizes.
In chapter 7, conclusions are drawn and future directions are discussed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
1.3 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
T. Plowman, W. Reichert, C. R. Peters, H. K. Wang, D. A. Christensen, J. N.
Herron, "Femtomolar sensitivity using a channel-etched thin film waveguide
fluoroimmunosensor," Biosensors & Bioelectronics, pp. 149-160,1996.
E. Kress-Rogers, "Handbooks o f biosensors and electronic noses: medicine,
food and the environment," 2nd ed: CRC press, 1997.
G. T. Hermanson, Bioconjugate Techniques. San Diego: Academic Press,
1995.
G. J. Sprokel, J. D. Swalen, "Chapter 4 The Attenuated Total Reflection
Method," in Handbook o f Optical Constants o f Solids II, E. D. Palik, Ed.:
Academic Press, 1991.
W. D. Wilson, "Analyzing biomolecular interactions," Science, vol. 295, pp.
2103-2105,2002.
K. Johnston, M. Mar, S. Yee, "Prototype o f a multi-channel planar substrate
SPR probe," Sensors and Actuators B, vol. 54, pp. 57-65, 1999.
I. Willilams, T.A. Addona, "The integration o f SPR biosensors with mass
spectrometry: possible applications for proteome analysis," Biotechnology,
pp. 45-48, 2000.
http://www.biacore.com/lifesciences/index.html.
J. R. Krone, R.W. Nelson, D. Dogruel, P. Williams, R. Granzow,
"Interfacing biomolecular interaction analysis with mass spectroscopy,"
Analytical Biochemistry, vol. 244, pp. 124-132, 1997.
P. S. Bernard, B. E. Pennock, H. P.Schwan, "Further observations on the
electrical properties o f hemoglobin-bound water," Journal o f Physical
Chemistry, vol. 73, pp. 2600-2607, 1969.
C. G. Essex, M.S. Symonds, R.J. Sheppard, E.H. Grant, R. Lamote, F.
Soetewey, M.Y. Rosseneu, H. Peeters, "Five-component dielectric
dispersion in bovine serum albumin solution.," Physics in Medicine &
Biology, vol. 22, pp. 1160-1167, 1977.
E. H. Grant, R.J. Sheppard, G.P. South, Dielectric behavior o f biological
molecules in solution, 1 ed: Oxford University Press, 1978.
S. Mashimo, N. Miura, N.Shinyashiki, T. Ota, "Dielectric study on
molecular motions o f poly(glutamic acid) in aqueous solutions over a
frequency range o f 10e5~10el0 Hz," Macromolecules, vol. 26, pp. 68596863,1993.
S. Bone, "Dielectric studies o f native, unfolded and intermediate forms o f
beta-lactamase," Physics in Medicine & Biology, vol. 39, pp. 1801-1809,
1994.
M. Suzuki, J. Shigematsu, T. Kodama, "Hydration Study o f Proteins in
Solution by Microwave Dielectric Analysis," Journal o f Physical Chemistry,
vol. 100, pp. 7279-7282, 1996.
T. Kamei, M. Oobatke, M. Suzuki, "Hydration o f apomyoglobin in native,
molten globule, and unfolded states by using microwave dielectric
spectroscopy," Biophysical Journal, vol. 82, pp. 418-425, 2002.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
[17]
Y. Hayashi, N. Shinyashiki, S. Yagihara, "Dynamical structure o f water
around biopolymers investigated by microwave dielectric measurements
using time domain reflectometry method," Journal o f Non-Crystalline Solids,
vol. 305, pp. 328-332, 2002.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
Chapter 2 Detection of Biomolecular Interactions
For this research project, we focus on developing a new signal detection
technique, which is to detect BMIs through sensing the dielectric property changes at
the aqueous/solid interface in the microwave regime. Research has shown that
biomolecules have distinctive dielectric features in the microwave range [1-4],
whereas in optical domain, the refractive index o f proteins is relatively constant [5].
This suggests that the dielectric technique could be superior to the optical sensing
techniques since it may be possible to gain molecular structural information as well
as detecting the binding events. The dielectric property involved with BMIs have not
been extensively studied or fully understood mainly due to the fact that detection
techniques geared for detecting the BMIs in vivo are not available, which motivates
our interest in developing a new BMI sensor. Our work focuses on studying o f the
BMIs at the transmission line solid/liquid interface, and obtaining a fundamental
understanding o f what gives rise to these dielectric changes, both from the
electromagnetic and molecular biology point o f view. We will incorporate the
available surface biomodification techniques and developed fluidics handling
schemes to build up our new sensor system.
2.1 Anatomy of Biomolecular Interaction Sensors
A surface sensitive biosensor is an analytical device that integrates a
biological element on a solid-state substrate, and allows a specific biomolecular
interaction with the analyte. The biological element is a layer o f molecules applied
for biorecognition, such as enzymes, receptors, peptides, single stranded DNA, or
even living cells. The biological element is in contact with a signal transducing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
element that converts the analyte-receptor reaction into a quantitative electrical or
optical signal. A transducer should be highly specific for the analyte o f interest. It
should also be able to respond in an appropriate concentration range and have a
moderately fast response time.
Current transducer techniques can be divided into two categories: molecular
label techniques and label-free techniques. Molecular labeling uses molecular tags
(fluorescence, or radioactive labels) conjugated to one or several interacting
molecules. The fluorescence detection technique is by far one o f the most sensitive
detection methods with the sensitivity reaching to femtomolar concentration [6]. But
the molecular labeling techniques suffer some limitations as described in Chapter 1.
Label free detection methods eliminate those disadvantages, which makes them very
attractive even though to date their sensitivities do not match the labeling technique.
Surface plasmon resonance (SPR) technique is one of the most widely used
label-free sensing technique for measuring binding events in-situ in physiologic
relevant environments [7]. SPR occurs in the optical domain because the real
permittivity o f such metals is negative in the optical frequencies. The incident angle
( 0 ) at which the surface plasmon is excited depends on the refractive indices on both
sides o f the metal film. SPR is sensitive to refractive index changes in close vicinity
of the interface since the electromagnetic field is evanescent and decays rapidly. The
sensitive region is less than a wavelength in the bulk sensing media (usually about
200nm from the sensor surface). A BMI event perturbs a surface electromagnetic
mode that can exist at the interface between a noble metal (such as gold) and a
dielectric medium. The detection limit o f SPR is able to reach at least 5pg/mm2 [8],
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
which is sufficient for BMI studies even though not as good as the fluorescence
technique.
The SPR sensor has several limitations. One is that it indirectly measures the
change in the propagation constant by observing subtle shifts in intensity o f the light
reflected, i.e. it is not specific to the type o f ligand on the surface. Another limitation
is that the optical and geometric characteristics o f the system are not best suited for
the study o f binding events in mixed media (for instance in the presence o f a
nanoporous film or gel) because o f the tight coupling between the signal transduction
and binding interfaces. Finally, gold surface is commonly used in SPR. The gold
surface modification usually involves use o f the thiol ligands for biomolecular
immobilization. Thiol ligands are not chemically stable, which could interfere with
biologically relevant reactions.
SPR performs sensitive in-situ measurement at aqueous-solid interfaces
through measuring the dielectric property (refractive index) change caused by BMIs
in the optical frequency range. Changes in the dielectric properties o f aqueous
biological solutions in microwave regime are even larger than other parts o f the
electromagnetic spectrum. For instance, the dielectric permittivity o f water changes
significantly from about 80 at about 100MHz to a little below 20 at 35GHz [9] and
water is the major solvent in most biosamples. For this reason, developing a
dielectric microwave sensor is very attractive.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
2.2 Basic Permittivity Concepts
2.2.1 Polarization and Polarizability of the Material
There are polar and non-polar dielectric materials. In a polar dielectric
molecule, the center of positive charge is displaced from the center o f negative
charge. The value o f the dipole moment o f an uncharged molecule is given by total
positive or negative charge multiplied by the separation of the two centers o f charge.
A polar molecule possesses a permanent dipole moment even in the absence o f any
applied electric field. A non-polar dielectric molecule does not have a dipole moment
unless it is in the presence o f an electric field. The electric field induces the
molecular dipole moments by perturbing the electron cloud around the nucleus so as
to produce a separation of the center o f negative charge and the center o f positive
charge.
The polarization P of a dielectric material inserted between the parallel plates
o f a capacitor is equal to the electric flux density ( D ) on the plates minus the charge
density required to maintain the potential difference between the plates.
P = D - e0E
= £oSrE - e 0E = e0(£r - l ) E = £0z E
(2.1)
Z = er - I
where E is the E-field strength and %is called the relative dielectric susceptibility.
The polarization o f dielectrics arises from the finite displacement o f charges or
rotation o f dipoles in an electric field. At the molecular level, polarization involves
either the distortion o f the distribution o f the electron cloud within a molecule or
dipole moment rotations.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
2.2.2 Dielectric Relaxation and Complex Permittivity
The permittivity, often called the dielectric constant, is a complex number
and varies with frequency. In general, permittivity can be divided into molecular,
ionic and electronic contributions. At low frequencies in microwave regime, the
permittivity characterizes the atomic and molecular dipole moments o f a material.
For instance, a polar liquid such as water, e is greater than the square o f the optical
refractive index. The difference is due to reorientation o f the electric dipoles in the
applied field.
s = e'-je"
(2.2)
E(t) = Eejmt
If the applied E-field is
_
,------------------ _
(2.3)
D(t) = £„{£'-js " )E e Jm' = £0( ^ ( £ ' f +(e")2 )EeJ{a,-,p)
where tan (p = s "/ s ' . A complex permittivity implies a phase lag o f cp between the
electric flux density D and the electric field. The real part of the permittivity is a
measure o f how much energy is stored in the material, and the imaginary part o f the
permittivity is a measure o f how dissipative or lossy a material is to an external field.
As the frequency increases, the dipole moments are just not able to orient fast
enough to keep alignment with the applied electric field, and the polarization fails
with the reduction of permittivity. It is referred to as a dielectric relaxation. The
Debye and Cole-Cole equations are commonly used to model the polar liquid
relaxations. Detailed Debye model description is included in the Appendix A.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
2.3 Dielectric Property of Biological Materials
Biological materials have characteristic dielectric signatures in the
microwave frequency range. Biological molecules are polar in nature. The molecular
dipole moment depends on the size of the molecule and on the distribution o f charge.
Dielectric studies provide information about protein size, shape, and the distribution
of charge in the protein[10-13].The aqueous biological solutions have several
relaxation frequencies, and the dispersion curve can be broken down into three
principle relaxation regions, a, /?, and y relaxations [11].
The a-relaxation is around several hundred Hertz, and is attributed to the
relaxation o f counter ions surrounding the charged cell membranes, or may be from
the migration o f ions through the holes in the membrane [14].
The ^-relaxation arises from distortions of cellular membranes and
macromolecules. It occurs from kHz to several hundred MHz. One type o f /?dispersion arises from in-homogeneity o f the material, and provides information o f
the structure and width of the cell membranes. When a heterogeneous medium is
placed in an electric field, charges will accumulate at the various boundaries
separating regions o f different permittivity and conductivity. It takes time for the
charges to reach equilibrium. This brings a variation o f permittivity with frequency,
which is referred to as the Maxwell-Wagner effect [11]. Details about the MaxwellWagner interfacial effect is discussed in the Appendix B. Another /^-dispersion is
from the rotation o f biological molecules but not from water. The /^-dispersion
corresponds to the orientational relaxation o f the whole protein dipole in water in the
range from 0.1 to 2MHz depending on the protein size [1,13].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
The y-dispersion is attributed to the relaxation o f the water molecules. The
permittivity o f pure water has been measured with the relaxation frequency at around
17GHz, corresponding to a relaxation time o f 9.3ps. The low frequency permittivity
o f water is close to 80, and drops to a little below 20 at frequencies beyond 35GHz
[9].
For a biological aqueous solution, the permittivity at low frequencies is less
than 80 by an amount which is roughly proportional to the volume fraction o f bio­
material present. This lowering o f the permittivity below the value o f pure water can
be used to obtain information concerning the water immediately adjacent to the
solute molecules [11]. The decrease o f permittivity o f biological solutions also
provides us foundations for the BMI sensing. The BMI will bring the biomolecules
concentrate at the solid/liquid interface, and the interfacial permittivity should be
lower than the bulk solution above the interface.
There is a small dispersion region between the /? and y regions. This small
dispersion region has been observed in several protein aqueous solutions and referred
as 8 dispersion [1, 10,11,15, 16]. It has been proposed that 8 dispersion comes from
the relaxation o f the water molecules in the immediate proximity o f the biological
macromolecules. The water molecules are influenced by the strong electric fields
produced by the macromolecules close to them. Consequently, the relaxation time o f
such water is longer than that o f the pure liquid water.
The dielectric transmission line BMI sensor will provide a tool for studying
the biomolecular interactions as well as biomolecular-water interactions. Studying
the water relaxations ( y, S ) is o f great interest to biologists since as the natural
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
solvent o f biological macromolecules, water influences many aspects o f biological
function. Water-protein interaction has long been viewed as a major determinant o f
chain folding, conformational stability, internal dynamics, and binding specificity o f
globular proteins [17]. For instance, water bound to the protein can reduce the
electrostatic forces between polar species thereby facilitating structural mobility.
Water molecules compete with other intra-molecular interactions between polar
groups o f protein. Water also takes part in hydrophobic interactions with non-polar
amino acid side groups, making the protein interior a more energetically favorable
environment for these species [17].
2.4 Transmission Line Dielectric Measurement Techniques
Transmission line dielectric measurement techniques measure the
propagation constant y from which the complex permittivity e can be determined.
y = a + j P = k04 s
(2.4)
a is the attenuation factor and /? is the propagation phase constant.
If the dielectrics surrounding the conductor are not homogeneous, for
example, a planar transmission line with semiconductor substrate at one side o f the
conductor, and air at the other side, the dominant propagation mode is a quasi-TEM
mode. The propagation constant then becomes y =
k0, where s eff is called the
effective dielectric constant. The effective dielectric constant o f a transmission line is
frequency dependent, and is a function o f the transmission line geometry, and the
dielectric property of the transmission line substrates surrounding the conductor.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
The dispersion o f a transmission line is the frequency dependence o f the
propagation characteristics. The dispersion property is revealed as group velocity
and phase velocity o f an electromagnetic wave traveling through the transmission
line. The group velocity o f the propagation mode is proportional to l / ^ £ ^ , where
seff is the effective permittivity o f the waveguide.
2.4.1 Coaxial Transmission Lines
Coaxial transmission line apertures are widely used for dielectric
measurements over the frequency range o f 50MHz to around 12GHz. The dominant
propagation mode is TEM in a coaxial transmission line. Two types o f coaxial cell
configurations are commonly used. One type o f coaxial cell (Figure 2.1 (a)) consists
of a coaxial line, i.e. a hollow metal cylinder with a concentric inner conductor. At
the end o f such a line is a flat metal plate which acts as a short circuit. The liquid or
solid sample (usually in powder) is filled in the cell. An electromagnetic wave
travels through the sample and is reflected by the short circuit. The reflected wave
will interfere with the incident wave so that a standing wave is set up within the
sample. The permittivity of the sample determines the shape and size o f the standing
wave. It can be calculated from the measured standing wave parameters.
Another configuration (Figure2.1 (b)) is a 5 0 0 impedance transmission line
with a flat end, for instance the Agilent 85070D dielectric probe. The probe is
directly immersed into the liquids or semi-solids to measure the relative permittivity.
The probe transmits a signal into the material under test. The measured reflected
response from the material is then related to its dielectric properties. Methods for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
extracting the dielectric constant from the coaxial transmission line measurements
are well developed [18,19].
The coaxial transmission line is designed for measuring bulk liquid or solid.
From this point o f view, the planar transmission lines are more suitable for dielectric
measurement from BMIs as oppose to the coaxial transmission line. They provide a
planar surface for bio-ligand modifications, and biomolecular interactions. Besides,
they are easy to couple with fluidic systems. Most importantly, the electromagnetic
fields are highly concentrated at the planar transmission line interface, which makes
it very efficient to couple to the dielectric behavior o f BMIs.
Outer Conductor * Metalic Disc
forming short
Inner Conductor
Dielectric Sample
(a)
Outer Conductor-*""
Inner Conductor*
50 Ohm with flat end
(b)
Figure 2.1 Sam ple cells used w ith coaxial transmission line to measure dielectric constant o f
liquids a n d solids, (a) is coaxial transmission line w ith a m etallic disc fo rm in g short (b)is a
coaxial transm ission line with 50 ohm fla t end.
2.4.2 Planar Transmission Line as Dielectric BMI Probe
Most recently, some research groups have started to study the dielectric
properties o f biological samples using planar transmission line configurations such
as stripline, microstrip line, and coplanar waveguides [20-22]. Hefti et al. published
their work on measurement o f the dielectric property o f large aqueous-based
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
molecules such as proteins and nucleic acids in microwave frequency range using a
stripline configuration [21]. They modified the center conductor with biomolecules.
It is claimed that the device could identify the molecular stmctures, changes in
structures, and various types o f intermolecular interactions via the direct detection of
the dispersion properties o f the system. Shortly after Hefti’s report, Facer and Sohn
reported a coplanar waveguide device to perform dielectric spectroscopy study on
bio-aqueous samples confined to a microfluidic channel. Measurements were made
for a broad frequency range from 40Hz to 26.5GHz [22]. The dielectric properties of
bulk biological liquid samples were measured instead o f the biomolecular
interactions at the transmission line surface. In microwave frequency regime,
distinguishable spectral responses were observed for several biological aqueous
solutions including DNA solutions, hemoglobin solution and Escherichia Coli (E.
Coli.). A decrease in power dissipation was observed within the sample compared
with the pure saline solution, which represented an increase in the transparency of
the medium to microwaves.
Hefti and Sohn’s work demonstrated that the planar transmission line
structures could be used to measure the dielectric property of different biological
solutions. A variety o f biomolecules showed different qualitative responses in
microwave regime, to which the authors referred as ‘signatures’. Their results
indicated that aqueous bio-solutions were better dielectrics (lower permittivity) than
water alone at the frequency, which suggested that BMIs occurring at the
transmission line interface will lower the dielectric constant.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
We selected a coplanar transmission line (CTL) configuration for in this
thesis project. A CTL consists o f a strip o f thin metallic film on the surface o f a
dielectric planar substrate with two ground electrodes running adjacent and parallel
to the signal strip. The dominant propagation mode on CTL is quasi-TEM [23, 24],
For our CTL device, we chose a glass substrate, and the metal strip was gold. Refer
to Figure 2.2 for details. The use o f a coplanar waveguide has several advantages for
liquid measurements. First, the coplanar sensors can be placed conveniently in
contact with the test material and therefore provide non-destructive measurements to
test dielectric liquid sample. Second, the CTL configuration provides a planar glass
surface for biomolecular immobilization. There are well-developed protein
deposition protocols for a glass substrate with good stability. As opposed to Hefti’s
stripline configuration which requires biomodification on the gold conductor strip,
our CTL provides a glass surface between the signal and ground strip for bio-ligand
molecule deposition. The thiol ligand which is commonly used for gold, is not very
stable and may interfere with biological reactions. Third, it is easy to couple the CTL
structure with fluidic cavities on top to handle small volume fluid samples, which is
a common requirement in bioanalysis.
Signal strip
Ground strip 4—
strip
^
►Glass substrate
Figure 2.2 A cross-section view o f the coplanar transmission line
2.5 Planar Transmission Line Surface Biomolecular Deposition Schemes
The protein immobilization on a BMI sensor surface is an important step of
sensor implementation, since it will determine the function o f a BMI probe. There
are many well-developed protein immobilization protocols available for glass surface.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
This section serves as a brief overview of these surface functionalization and protein
immobilization techniques.
Advantages o f using glass substrate are its stable chemical property and low
cost. Most o f the protein immobilizations on glass substrates are based on protocol
chemistries that have been developed for DNA immobilizations. Glass slides need to
be pre-coated with a thin organic film before the immobilization o f proteins. Usually
used for pre-coating are the silane reagents terminated with nucleophiles such as
amines or thiols, or with electrophiles such as epoxide or aldehyde. The glass surface
can also be pre-coated by physisorption o f polyelectrolytes such as polylisines.
These thin organic films enhance the biocompatibility of the surface and protect the
proteins from denaturation and structural changes during the immobilization step.
The reactive coatings o f the glass surface are mainly based on self-assembly
techniques. Self-assembled monolayers (SAMs) o f alkylchlorosilanes or
alkyloxysilanes are commonly made. A great variety o f silanization conditions
pertaining to temperature and reaction time, which are also strongly influenced by
the stability of the terminal functional groups have been described in the literatures
[25].
Generally, the thin organic SAMs provide biocompatible interfaces during
the chemical coupling of proteins to the substrate, either via covalent binding,
complex coordination chemistry or molecular interactions. For the covalent chemical
coupling, proteins should have the following chemical function groups in the side
chains o f their polypeptide backbone:-SH (Cysteine),-NH 2 (Lysine), -COOH ,OH(serine), Ph-OH. In principle, all o f such proteins can be used during the direct
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
chemical coupling reaction on specifically prepared SAM surfaces. In the following
chapters, we will describe the surface functionalization in details using epoxy- or
aldehyde terminated alkylsilanes or polysine to immobilize antigens such as protein
A, or biotin, and antibodies such as immunoglobulin.
2.6 Conclusions
We will focus on detecting the dielectric property change from BMIs at the
planar transmission line interface. A CTL configuration is chosen for building the
BMI sensor system. The well developed surface modification methods and
bioconjugate techniques make it possible for us to immobilize the ligand molecules
on the CTL probe surface and design the experiment to detect the specific BMIs at
the solid/liquid interface. Development of such technique is o f great scientific
significance. The dielectric permittivity study characterizes the molecular dipole
moment changes, which will help us better understanding the biomolecular
conformation change, and water-biomolecular interactions during BMIs in a wellcontrolled manner at the interface.
Detecting the interfacial reaction through permittivity change is very
challenging. It requires measurement o f the permittivity changes occurring within a
few nanometers at the solid/liquid interface using microwave signal with the
wavelength on the order of millimeters. Hefti and Sohn’s work suggested that BMIs
occurring at the transmission line interface will lower the dielectric constant. The
feasibility test experiments in the following chapter will answer the question that if it
is possible to measure such permittivity changes occurring within a few molecular
layers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
2.7 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
H. P. Schwan, K. Foster, "RF-Field interactions with biological systems:
electrical properties and biophysical mechanisms," Proceedings o f the IEEE,
vol. 68, pp. 104-113,1980.
J. Bateman, C. Gabriel, G. Evans, E. H. Grant, "Dielectric properties o f urea
and avetamide in aqueous solutions," Journal o f Chemical Society o f
Faraday Transactions, vol. 86, pp. 321-328, 1990.
S. Mashimo, N. Miura, N.Shinyashiki, T. Ota, "Dielectric study on
molecular motions o f poly(glutamic acid) in aqueous solutions over a
frequency range o f 10e5~10el0 Hz," Macromolecules, vol. 26, pp. 68596863,1993.
S. Bone, "Dielectric studies o f native, unfolded and intermediate forms of
beta-lactamase," Physics in Medicine & Biology, vol. 39, pp. 1801-1809,
1994.
H. Elwing, "Protein absorption and ellipsometry in biomaterial research,"
Biomaterial, vol. 19, pp. 397-406, 1998.
T. Plowman, W. Reichert, C. R. Peters, H. K. Wang, D. A. Christensen, J. N.
Herron, "Femtomolar sensitivity using a channel-etched thin film waveguide
fluoroimmunosensor," Biosensors & Bioelectronics, pp. 149-160, 1996.
W. D. Wilson, "Analyzing biomolecular interactions," Science, vol. 295, pp.
2103-2105,2002.
E. Kress-Rogers, "Handbooks o f biosensors and electronic noses: medicine,
food and the environment," 2nd ed: CRC press, 1997.
C. H. Collie, J.B. Hasted, D.M. Ritson, "The dielectric properties o f water
and heavy water," Proceedings o f the Physical Society, vol. 60, pp. 145-160,
1948.
C. G. Essex, M.S. Symonds, R.J. Sheppard, E.H. Grant, R. Lamote, F.
Soetewey, M.Y. Rosseneu, H. Peeters, "Five-component dielectric
dispersion in bovine serum albumin solution," Physics in Medicine & Biology,
vol. 22, pp. 1160-1167, 1977.
E. H. Grant, R.J. Sheppard, G.P. South, Dielectric behavior o f biological
molecules in solution, 1 ed: Oxford University Press, 1978.
S. Bone, B.Z. Ginzburg, H. Morgan, G. Wilson, B. Zaba, "Time-domain
dielectric spectroscopy applied to cell suspensions," Physics in Medicine &
Biology, vol. 38, pp. 511-520, 1993.
M. Suzuki, J. Shigematsu, T. Kodama, "Hydration Study o f Proteins in
Solution by Microwave Dielectric Analysis," Journal o f Physical Chemistry,
vol. 100, pp. 7279-7282, 1996.
C. G. Essex, G.P. South, R.J. Sheppard, E.H. Grant, "A bridge technique for
measuring the permittivity o f a biological solution between 1 and 100 MHz,"
Journal o f Physics E - Scientific Instruments, vol. 8, pp. 385-389, 1975.
J. B. Bateman, G. F. Evans, P.R. Brown ,C. Gabriel, E.H. Grant, "Dielectric
properties o f the system bovine albumin : urea : betaine in aqueous solution,"
Physics in Medicine & Biology, vol. 37, pp. 175-182, 1992.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
S. Bone, "Dielectric and gravimetric studies o f water binding to lysozyme,"
Physics in Medicine & Biology, vol. 41, pp. 1265-1275, 1996.
B. Alberts, Molecular biology o f the cell, 3rd ed. ed. New York: Garland Pub.,
1994.
R. Cole, "Evaluation o f dielectric behavior by time domain spectroscopy
(complex permittivity)," Journal o f Physical Chemistry, vol. 79, pp. 14691474,1975.
R. Cole, S. Mashimo, P. Winsor, "Evaluation o f dielectric behavior by time
domain spectroscopy 3. precision difference methods," Journal ofPhyscial
Chemistry, vol. 84, pp. 786-793, 1980.
S. S. Stuchly, C.E. Bassey, "Microwave coplanar sensors for dielectric
measurement," Measurement Science and Technology, vol. 9, pp. 1324-1329,
1998.
J. Hefti, A. Pan, A. Kumar, "Sensitive detection method o f dielectric
dispersions in aqueous-based, surface-bound macromolecular structures using
microwave spectroscopy," Applied Physics Letters, vol. 75, pp. 1802-1804,
1999.
G. R. Facer, D.A. Notterman, L.L. Sohn, "Dielectric spectroscopy for
bioanalysis: from 40Hz to 26.5GHz in a microfabricated wave guide,"
Applied Physics Letters, vol. 78, pp. 996-998, 2001.
C. P. Wen, "Coplanar waveguide: a surface strip transmission line suitable
for reciprocal gyromagnetic device applications," IEEE Transaction on
Microwave Theory and Techniques, vol. 17, pp. 1087-1090, 1969.
K. C. Gupta, R. Garg, I.J. Bahl, Microstrip lines and slotlines, 1st ed: Artech,
1979.
A. Ulman, "Formation and structure o f self-assembled monolayers,"
Chemical Reviews, vol. 96, pp. 1533-1554, 1996.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
Chapter 3 Coplanar Transmission Line Probe Implementation and
Detection of Biomolecular Surface Interaction
3.1 Introduction
This chapter focuses on the coplanar transmission line (CTL) biomolecular
interaction (BMI) probe implementation and the feasibility study. The CTL probe is
a new sensing method to detect changes occurring at the interface. The glass region
between the signal and ground strip is chemically modified to immobilize ligand
molecules for the purpose to detect the target molecules in a liquid sample. A
specific BMI reaction takes place at the aqueous-solid interface that changes the
permittivity. ‘Is it possible to detect such BMI reaction that occurs within several
molecular layers, with an electrical microwave measurement?’ This important
question is answered in this chapter with affirmation.
Two sets o f feasibility experiments were performed. The first set of
feasibility experiments demonstrated that the biosensor technique could detect a
specific BMI at the transmission line solid/liquid interface. Specific BMIs were
detected with the level o f detection on the order o f pg/ml. The reactions involved
biotin to anti-biotin, biotin to avidin, and antibody rabbit y- immunoglobulin (R-IgG)
to anti-R-IgG bindings.
The second feasibility experiment was a real-time measurement that follows the
affinity interaction between protein A (PA, MW 42,000) and y-immunoglobulin (IgG,
MW~150,000) [1, 2]. The PA was immobilized on the CTL surface. We determined
the sensitivity, dynamic range of the CTL probe, and finally the affinity constant was
estimated for this BMI reaction.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
3.2 Coplanar Transmission Line Probe
The new BMI probe is based on a CTL. The following sections describe the
CTL design and fabrication, the microfluidics which is coupled with the transmission
line to form the reaction cavities, and the time domain measurement set-up.
3.2.1 Coplanar Transmission Line/ Microfluidics System Design and
Fabrication
In a conventional transmission line design, the characteristic impedance o f
the CTL should be equal to the input impedance o f the detector instrument. However,
it is impossible to design a matched CTL for our case. The aqueous biological
solutions on top o f the transmission line are very dispersive. This is mainly due to
the dielectric constant of water which changes from 80 to a little below 20 over this
frequency range (1M Hz- 35GHz) [3]. We simply chose a CTL with 50Q
characteristic impedance using a glass substrate, assuming air on top instead o f water.
This design provides us a well defined configuration for the purpose o f experimental
confirmation.
The CTL probe was modeled as a three-layer CTL structure (Figure 3.1).
The CTL characteristic impedance design was based on quasi-static analysis which
assumed the propagating mode in the transmission line was a pure TEM mode at
zero frequency. The impedance o f the CTL was determined by the dimensions o f the
signal strip s, the gap between the signal and ground strips g, and the permittivity o f
the dielectrics below and above the CTL. The analytical equations for calculating the
characteristic impedance are well documented [4-6].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
PMMA
g=50um
s=400um
Glass
Figure 3.1 Three-layer coplanar transmission line (CTL) configuration with (1) glass substrate
( £ r = 4.6) (2) a thin aqueous sam ple layer (assum ed to be air f o r CTL design) a n d (3) the PMMA
sam ple cover ( £ r = 2 .7 ).
We designed the CTL with the signal strip width s = 400pm, and the gap
between the signal and ground strips g = 50pm. The quasi-static design result was
confirmed by measuring the frequency response (500MHz to 20GHz) o f the CTL
using a vector network analyzer (Agilent 8703A). The frequency response (S 2 1 ) o f a
2cm long CTL with designed dimensions, and air on top had less than -lOdB loss up
to 15GHz (Figure 3.2). The experimental result confirmed that a pulse with a
bandwidth o f 15 GHz could propagate through the probe in the absence o f liquid on
-3-
-6 -
CO
-9-
-12-
Frequency re sp o n se (S 21) of CTL probe with air on top
-15
0
2
4
6
8
10
12
14
16
18
20
Frequency (GHz)
Figure 3.2 Frequency response (S21) o f the CTL p ro b e with air on top, m easured with the Agilent
8703A N etw ork Analyzer.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
A 2cm long CTL was fabricated on a 5cm by 2.5cm borosilicate glass
substrate. A 0.5nm Ti was first deposited as an adhesion layer, followed by a 500nm
thick gold film deposited using an electron-beam evaporator (SEC600, CHA
Industries, Fremont, CA, deposition rate 0.7nm/s, base pressure 5x1 O'6 Torr, at room
temperature). The gold layer was deposited immediately after the Ti layer without
breaking the vacuum to the system.
The ground and signal electrodes were photo-lithographically defined. First,
a positive photoresist (S-1813, Shipley Co.) was spin coated (3000rpm for 30
seconds) and baked on a hot plate for 1 minute. The substrate was aligned with a
photomask and exposed to the UV light for 3 seconds. The exposed photoresist in the
gap regions between the signal and ground strips was washed away with M F-319
developer (Shipley Co.). The underlying gold was removed using Aqua-Regia (a
mixture o f hydrochloric acid and nitric acid). Finally, the Ti adhesion layer
underneath was removed by dipping in a buffer oxide etch (1:1 N H 4 F and HF) for 10
seconds. The remaining photoresist was removed with a photo-resist stripper (PRS3000, Baker Co.) followed by rinses in isopropanol, and de-ionized water. The
patterned substrate was then dried with nitrogen.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
Ludte Cover V\flth Fluidic Inlets
nr r
ii
i i
ii
ii
ii
ii
ii
ii
ii
ii
ii
ii
ii
Reference Channel
i
i
i
i
i
i
i
i
i
i
i
i
i
i
' 1"T ~ .....
1 1
1 1
1 1
1 1
1 1
■ ■
Sample CHShri
Y
9 9 99
Target Molecules
Y
Ground Electitxle
Signal Electrode
Ground Electrode Signal Ele
Ground Electrod*
►Ligand Molecules
G lass gap region in both channels w ere modified
ligand molecules
Figure 3.3 A cross section o f the CTL probe. Ligand protein molecules w ere im m obilized onto the
glass interface between the signal and ground electrode. Target molecules flo w through the flu id ic
channel a n d bin d selectively to the ligands, which lea d to a change in permittivity.
Figure 3.3 shows details about the probe configuration. Two identical CTLs
were fabricated on a glass substrate in order to minimize pulse jittering and signal
drifting. One formed the reference channel, and the other formed the sample channel.
A 2cm long fluidic chamber was placed on top o f each transmission line. Channels
and cavities were machined into a poly (methacrylyte) (PMMA) cover that in
combination with gaskets formed a liquid chamber [4mm (w) x 20mm (/) x 200pm
(h, approx.)], holding approximately 10 pi liquid over each transmission line. The
fluidic channels in the PMMA cover were fitted with chromatography flangeless
fittings and connected to 1/16 OD tubing (Upchurch Scientific). The liquid was
delivered with a syringe pump (Harvard Apparatus, Holister, Mass.).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
3.2.2 Time Domain Transmission System Set-up
We chose a time domain transmission (TDT) measurement over a frequency
domain measurement for several reasons. First was to avoid the impedance mismatch
termination problem. Since the impedance o f the CTL with aqueous solution on top
changes over the microwave frequency range, the impedance mismatch problem will
plague the frequency domain measurement. In a time domain measurement, the
reflections from the impedance mismatch can be eliminated by time- windowing.
Second, for a frequency domain measurement, the CTL probe configuration requires
a complicated de-embedding procedure to define the test plane. But in the time
domain, the permittivity change at the interface can be easily measured by
comparing the change in the transit time o f a fast pulse transmitted through the CTL.
The TDT measurements were performed with a high-speed digital sampling
oscilloscope (HP54750A). A 50kHz pulse train generated by an 18GHz time-domain
reflectometry module (HP-54754A) triggered a pulse generator (Picosecond Labs,
PSPL4015C) to create step pulses that were then differentiated by a pulse forming
network (PSPL5208). This step pulse created -3 volts impulses (full-width half
maximum= 1 lps). The time jittering was smaller than 1.8ps. The pulse shapes were
digitized by a 50GHz high speed module (HP54752A) using a lOOOps sampling
window with 4096 sampling points. For each measurement, 32 impulses were taken
and averaged.
To avoid the room temperature fluctuation, we kept the whole fluidic fixture
on a temperature-control stage with the temperature set at 30 °C. The temperature
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
was monitored by a thermistor (Mcshane Inc. TS67-170). The temperature variation
was within ±0.1 °C throughout the experiments.
3.3 Feasibility Test Experiments
3.3.1 Feasibility Series 1- Specific Binding Experiments: Biotin to Anti-biotin, and
Rabbit y-Immunoglobulin (IgG) to Anti-rabbit IgG
The goal o f the first series o f feasibility tests was to see if the CTL probe can
detect a BMI at solid/liquid interface on the transmission line. The specific
biomolecular interactions between ligand and target molecules changed dielectric
properties, which were detected as a pulse shape and a transient pulse change. Two
binding assays were tested. First assay was anti-biotin (target) and avidin (target)
binding to biotin (ligand). Second assay was anti-rabbit IgG (target) to rabbit IgG
(ligand). We measured the pulse transmitted through the probe before and after the
bio-binding reactions.
3.3.1.1 Experimental
Biotin or rabbit IgG ligand molecule was immobilized in the glass regions
between the signal and ground strip. The glass surface was precoated with
glycieoxyproply-trimethoxysylane (GOPS, Gelest SIG5840). The GOPS film is an
epoxide-rich surface that reacts slowly with nuleophilic(amines and thiols) making
stable bonds with the proteins. In a typical procedure, the substrates were subjected
to oxygen plasma cleaning (5 minutes at a power o f 50W with an oxygen pressure o f
40mTorr). Then, the substrate was immersed in 1% GOPS methanol solution
overnight. After the substrate was then rinsed with copious amount of methanol to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
remove the excess GOPS, the surfaces were annealed in an oven at 120 °C for 30
minutes.
The biotinylated surfaces were made by incubating the GOPS substrates with
fucossylamide-biotin labeled bovine serum albumin (B-BSA) (Pierce 21346). The
Rabbit-IgG (R-IgG) surface was prepared by incubation for 30 minutes with R-IgG
(Sigma 15006) solution with a concentration o f 10~15pg/ml. Excess B-BSAs or RIgGs were then removed by rinsing with water. The remaining active non-specific
sites were blocked by incubation with 1% BSA solution in phosphate buffer saline
(PBS). All chemical reagents were used as received.
In all cases the surface immobilization processes were verified by using
fluorescence labeled (Cy3) antibodies on control glass slides. The control slides were
glass slides exposed to the same protein immobilization procedure as the CTL probe.
For the biotinylated surface, the extrAvidin-Cy3 conjugate (Sigma E4142, F/P= 4.8)
was used for verification, and for the rabbit IgG surface, the sheep anti-rabbit IgGCy3 conjugate (Sigma C2306 F/P=3.6) was used. Cy3 labeled proteins emit the
fluorescence at the wavelength o f 530- 560 nm. A fluorescence image (Axon
Instruments GenePix 4000A, PMT=500 V) was taken on the control slides to
evaluate the protein immobilization yield. Scans resulted in signal counts o f the order
o f S>10,000 -30,000 and background o f the order o f 100-200 counts. From our
experience, this S/B ratio range o f 100-200 is a typical indication o f successful
immobilization o f protein monolayer.
The binding experiment on a biotinylated CTL probe was performed as
follows. First, the reaction chamber was filled with the PBS buffer. A transmitted
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
pulse was recorded as a reference. Then, the monoclonal anti-biotin mouse IgG
(Sigma 5585) (4pg/ml) was injected into the reaction channel. After thirty minutes,
the reaction channel was flushed with PBS buffer to get rid o f excess anti-biotin. The
formed biotin-anti-biotin complex remained on the CTL surface. The transmitted
signal trace was recorded. A decrease in transient time was observed, suggesting this
thin layer o f protein complex lowered the effective dielectric constant. The
extrAvidin (20pg/ml) was added to the reaction channel after the anti-biotin. After
thirty minutes, the reaction channel was washed with PBS buffer, and the signal trace
was recorded and compared with the reference.
For the second bio-assay experiment, the CPW transmission line gap surface
was modified with rabbit IgG to detect the anti-rabbit IgG in the liquid sample. This
experiment was performed to confirm the first experiment and compare the CTL
response to different binding assays. The anti-rabbit IgG concentration was
2.5pg/ml. Anti-biotin mouse IgG did not bind with the rabbit IgG molecules
specifically and was applied as the non-specific binding reagent with a concentration
of 20pg/ml for the control experiment.
The pulse transmissions through PBS buffer in the reference and sample
channel were monitored for 2 hours before the binding reaction. The signal was
stable and variation was negligible (O .lp s).
3.3.1.2 Results
Figure 3.4(a) shows the pulse transmission though PBS buffer (reference) and
the transmitted pulse after the biotin to antibiotin binding. The pulse transit time was
reduced and the pulse shape changed. The pulse difference by subtracting the sample
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
pulse from the reference pulse is shown in Figure 3.4(b). Figure 3.4(c) shows the
pulse difference with avidin introduced after the anti-biotin. The biotin interaction
with avidin is among the strongest non-covalent affinities known [4]. Avidin can
displace the anti-biotin that has already bound to biotin, and then bind with the
exposed biotin on the surface. The result in Figure 3.4(c) indicated additional pulse
time shift after avidin was introduced.
0. 00
-
^
0.02
-0.04
oo>T
H
o
-0.06
>
-0.08
-
P B S buffer
a n t i - b i o tin
0.10
0
200
400
600
800
1000
Time (ps)
Figure 3.4(a) The p u lse transmission through PBS buffer before (dashed line) an d after (solid line)
the biotin-anti biotin binding reaction.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
1o
0.005
05
10
a n t i - b io tin
15
0
200
400
600
800
1000
Tim e(ps)
Figure 3.4(b) The transient pu lse difference from PBS buffer base line (dashed line) an d the transient
pu lse difference after the biotinylated surface exposed to anti-biotin.
10
0.005
0. 000
-0.005
10
a n t i - b io tin
15
0
200
400
Time
600
800
1000
( p s )
Figure 3.4(c) The transient p u lse difference o f the biotinylated surface exposed to anti-biotin (solid
line) an d avidin (dashed line).
Figure 3.5 shows the experimental results o f R-IgG to anti-R-IgG binding. Figure
3.5(a) shows the CTL modified with R-IgG exposed to the non-specific anti-mouse
IgG. The anti-mouse IgG does not bind with rabbit-IgG specifically; therefore, the
sensor shows low response to it (Figure 3.5(a)). Figure 3.5(b) shows the specific
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
binding o f rabbit IgG (ligand) with anti-rabbit IgG.
010
0.005
0.000
-0.005
010
015
P B S buffer
anti-m o u s e
0
200
400
600
800
IgG
1000
Tim e ( p s )
Figure 3.5(a) The transient pu lse difference o f PBS buffer base line (dashed line) a n d the p u lse
difference after the CTL surface -with R-IgG exposed to anti-mouse IgG(non-specific binding).
010
0.005
000
-0.005
010
015
P B S buffer
a n t i - r a b b i t IgG
0
200
400
600
800
1000
Tim e ( p s )
Figure 3.5(b) The transient pu lse difference o f PBS buffer base line (dashed line) an d the transient
tim e pu lse difference after the CTL surface with R-IgG exposed to anti-rabbit IgG.
3.3.1.3 Discussions
These feasibility experiments definitely demonstrated that the CTL sensor
system could perform very sensitive detection o f biomolecular interactions at the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
solid/liquid interface. Specific biomolecular bindings such as biotin-antibiotin,
biotin-avidin, and immunoglobulin (rabbit IgG and anti-rabbit IgG) were detected
with the detection level on the order o f pg/ml. The output pulse also showed
different pulse distortion features that could be indicative to specific binding
reactions. This suggested that the CTL probe might be able to differentiate different
BMI events.
The geometry o f the CTL concentrated the electric field in the glass gap
region between the signal and ground strip. The glass gap region is where the BMI
took place. When the water region with high permittivity at the transmission line
interface was displaced by the low dielectric protein layer, the effective dielectric
constant decreased, corresponding to the pulse propagating through faster (i.e. the
decrease o f pulse transit time).
In these BMI experiments, we focused on measuring the relative transmission
line dispersion changes instead of measuring the absolute changes o f the dielectric
property. It is important to point out that it was not necessary to extract the dielectric
constant for the detection o f a BMI event. Extracting the absolute dielectric constant
from the measurement o f the planar transmission line case is very challenging for the
two reasons. First, the dielectric property o f aqueous solutions is very temperature
sensitive, which means the dielectric property changes with even a slight temperature
variation. The absolute dielectric constant measurement requires accurate
temperature control that is very hard to implement for a normal lab. In a relative
measurement, we could have a reference channel to minimize the temperature
interferences. Second, we can not extract the dielectric constant without a well
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
developed waveguide model. The electric field distribution on top o f CTL is
inhomogeneous, which makes the electromagnetic properties o f the composite
transmission line very difficult to accurately model. For these reasons, we focused on
the measurements o f the transient time change, which correlated to the effective
dielectric constant (£e/f) changes.
The decrease o f the pulse transit time during the BMI correlated to a decrease
of the effective permittivity sejf. The pulse transit time is directly related the group
velocity o f the electromagnetic mode. The group velocity of the electromagnetic
mode is proportional to 1/
,
where £effis a function of the dielectric property
around the CTL transmission line. A BMI at the interface caused a decrease o f £efj
because biomolecules with a lower permittivity (sr 2 -5) effectively displace the
water molecules (£>-50) from the interface. Therefore, an increase in density or in
thickness o f biomolecular layer resulted in a decrease in eejj.
This series o f experiment also showed that a more robust connector assembly
was needed. We improved the connector assemble method for a more efficient signal
launching and a better signal to noise ratio. In feasibility series 1, the microwave
signal was launched through SMA connectors which were pressed down on the CTL
signal strip. We changed the connection to the following. The substrate was placed
on a copper plate and SMA connectors were soldered (In/Ag solder, IND#290,
Indium Co. o f America) to the patterned gold electrodes. The launched input pulse
showed a better signal to noise ratio. As shown in Figure 3.6.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
— s i g n a l l a u n c h e d w ith s o ld e r e d c o n n e c to r s
- S i g n a l l a u n c h e d with c o n n e c t o r p r e s s e d d o w n ( t e s t l )
0 .0 0
-
0 .0 1
-
0.02
>
<
D
U
>
(Q
-0 .0 3
o
>
-0 .0 4
-0 .0 5
-0 .0 6
0
100
200
300
400
500
600
700
800
900
1000
t i me / ps
Figure 3.6 The f a s t pu lse transm itted through CTL p ro b e with PBS buffer on top. The sign al launched
with SMA connectors p re sse d down on the CTL (doted line) vs. the sign al launched with SMA
connectors soldered on the CTL (solid line).
3.3.2 Feasibility Series 2: Kinetics Study and Affinity Constant Estimation o f
Protein A and Rabbit IgG
3.3.2.1 Introduction
Understanding biomolecular interactions in the context o f biological function
requires more detailed information than just the answer to ‘Do they bind?’ The
ability to describe BMIs in terms o f affinity and kinetics characteristics is vitally
important for understanding the BMI interactions. An ideal BMI sensor should be
able to measure the protein concentrations, assess the binding specificity,
characterize the binding interactions in terms o f affinity, rate constants, and provide
identifications. Towards these objectives, we proceeded our second series o f
feasibility experiments. Our goal was to test if the system could perform a real time
measurement and binding affinity estimation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
There are two binding sites on an antibody immunoglobulin (IgG) molecule.
The Fragment crystallizable (Fc) portion o f the antibody binds specifically with PA
[2, 5 ]. The Fragment antigen binding (Fab) portion o f the antibody can be used for
other antibody interactions (often referred as a secondary antibody). We made the
real time measurement on the interaction between PA and rabbit IgG on a PA
immobilized CTL surface. We also determined the sensitivity and dynamic range o f
the CTL probe.
3.3.2.2 Experimental
The glass gap regions, which straddle the CPW signal strip, were modified
with aldehyde methoxysilane (AMS) (PSX1050 ‘Bio-Conext’ from United Chemical
Technologies). PA (Pierce 21184) was covalently linked to this surface by reductive
animation. Reductive animation refers to the nucleophilic addition then oxidationreduction reaction o f aldehyde with l°amines to give imines or 2° amines to give
enamines. In a typical procedure, the substrates were subjected to oxygen plasma
cleaning (5 minutes 50W, 40mTorr PO 2 ) followed by submersion in 1% AMS
methanol solution overnight. Substrates were then rinsed with copious amounts o f
methanol to remove excess AMS, and the surfaces were annealed in an oven in air at
120 °C for 30 minutes. PA was immobilized by incubating the AMS surface with ca
lOOpg/mL PA solution in the presence of NaBFU [4]. The surface immobilizations
were verified by using a fluorescence-labeled (Cy3) anti-rabbit IgG (Sigma.C2306)
secondary antibody. Time domain transmission measurements were made with a
high-speed digital sampling oscilloscope (HP-54750A) with the same set-up as used
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
in feasibility series 1. In the second test series, the real time data points were
collected every 20 seconds. An average pulse shape was recorded using 32 traces.
We followed the change in transient time and the pulse shape as the BMI
reaction progressed. To quantify these parameters, we used the cross-correlation. The
cross correlation between signal fi(t) and/ 2 (f) is defined as
+00
+
+00
J /|(r )/,(/ + r)rfr = —
-Q O
(3. 1)
-o o
where F(co) is the Fourier transform off(t) , and * means complex conjugate.
The change in pulse propagation time A r corresponds to where the cross correlation
o f two pulses is the maximum. The correlation is large when the signal is similar to
the reference. Refer to Figure 3.7 for a typical cross correlation calculation result.
The cross-correlation ratio T is defined here as maximum of cross-correlation o f
f xand f 2at time tmax divided by the auto-correlation o f the reference signal / 2 at tmax.
r = [ / (0 ® f 2( / ) L _ / [ / , « ) ® A (0 j,„ _
(3.2)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
PBS + Rabbit IgG
0.8 M g /m l
-20
At
= -6.56 ps
0.8
>
E
(0
c
O)
CO
-40
0.6
-60
Phosphate Buffer Sa
(PBS, pH =7)
-80
0.2
-100
-400 -300 -200 -100
0
100
200
300
400
500
600
700
0
100
200
300
400
T im e /p s
T im e (ps)
(a)
(b)
Figure 3 .7 M easurement o f the pu lse shift (Axp ) from PA-R-IgG binding using cross-correlation:
(a )tw o rea l pu lses taken during the reaction, (b) cross-correlation result o f the tw o pulse.
During the experiment, the data was processed in real time to determine the
cross correlation ratio T and the time-delay A r . Before each binding experiment,
both channels were exposed to PBS for 1.5 hours to measure a baseline drift
variation. The difference between the transmitted pulses in the signal
channel f
s (t)
correlation f
\PBS and the reference channel
s(t)
|ms. ®f ref( t )
\pbs,
f ref (/)
was calculated from the cross­
which served as the baseline. This baseline A r was
negligible, with a variation (a) always smaller than 0.02ps. This baseline variation
corresponded to a system detection limit for resolving changes in pulse transit time
o f 0.06ps (3a).
3.3.2.3 Dielectric Measurement Results and Fluorescence Results
After the baseline variation has been measured, the sample channel surface was
exposed to bovine serum albumin (BSA) PBS solution at high concentration
(500pg/ml) to block the CTL surface non-specific binding sites (see Figure 3.8).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
BSA does not specifically bind to PA, but may be physically adsorbed onto the
surface. The BSA solution was introduced to the sensor at a constant flow rate o f
O.lml/min. for about 10 minutes, followed by a PBS rinse for 15minutes. The
transient pulse shift o f the CTL sensor to the BSA bulk solution was Axp = -1,7ps
with a =0.2ps. The pulse shift after exposure to BSA Axp = -0.50ps, with a =0.2ps.
The origin o f the higher signal variation (baseline 3a = 0.06ps) observed after BSA
was introduced has not been identified. This experiment showed the sensor response
to bulk dielectric changes (introduced by BSA in high concentration). It also showed
a low-level o f non-specific protein binding occurred, as most o f the BSA was rinsed
away.
0.5
PBS
PBS+BSA
PBS
0. 0
-0 .5
1 .0
1 .5
-
2.0
0
10
20
30
40
Tim e (m in u te )
Figure 3.8 CTL p ro b e real-tim e response to the exposure o f BSA (500pg/ml). The pu lse
transit tim e shift was calculatedfrom cross-correlation o f the pu lse through the sam ple an d
the reference p u lse through PBS.
Two CTL slides (a and b) modified with AMS-PA were prepared and tested
individually. The two slides had different PA loading density. The subsequent
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
fluorescence test showed substrate (a) had higher PA loading than substrate (b). The
flow o f R-IgG (Sigma 15006) solutions were introduced into the reaction chamber
with the increasing concentrations. R-IgG o f several different concentrations
(50ng/ml, 250ng/ml, 2.5pg/ml, 50(j.g/ml, and lOOpg/ml) were introduced to the
sensor consecutively. Each sample volume was 500pl. The liquid sample delivery
rate was O.lml/min. The probe surface was exposed to the sample with a specific
concentration for 10 minutes, followed by a PBS buffer wash for about 7 to 10
minutes before the next sample with a higher concentration was introduced. Figure
3.9 shows the results taken with substrate (a). The vertical marker lines on the
binding curve indicate the valve switching points. The probe surface was exposed to
R-IgG or PBS as noted on the plot. The pulse transient time decreased, which
correlated to the build-up of protein complex layer at the transmission line interface.
We estimate that the surface reached saturation after exposing it to lOOpg/mL R-IgG
(~670nM) since no change in pulse transient time observed after exposing to a higher
concentration o f R-IgG. At this saturation level, the total pulse shift was -12ps. The
second substrate (b) was exposed to R-IgG using the same procedure, yielding a total
pulse shift o f -9.5ps. The probe exhibited a similar kinetics response.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
14
1 0 0 iig/m I ( 6 7 0 n M )
50 ng/m I (335nM
12
2.5 ng/m l (16.7nM
10
8
PBS
PBS
2 5 0 n g/m I (1 .6 7 n M )
PBS
6
PBS
4
2
■PBS
R - lg G 5 0 ng/m I (0 .3 3 n M )
0
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
Tim e(m inutes)
Figure 3.9 Real-tim e binding curve o f R-IgG on a PA-AMS surface. The p ro b e w as exposed to R-IgG
or PBS alternatively. The vertical marker lines on the binding curve indicate the valve switching
points.
To confirm the PA-RIgG binding took place, the slides were exposed to a
secondary antibody, Cy-3-labeled anti-rabbit IgG (Sigma.C2306). The R-IgG bound
with PA on its Fc portion, its Fab (Fragment antigen binding) portion was still
available to bind with the fluorescence labeled anti-rabbit IgG. Fluorescence image
profiles o f Cy3 were obtained using an Axon scanner (Axon Instruments GenePix
4000A). The fluorescence distribution profile at a cross section o f the substrate (a) is
shown in Figure 3.10. The two feature peaks with high fluorescence intensity
confirmed that R-IgG binding was predominantly inside the glass regions. Under
present experimental conditions, we did not observe substantial quenching o f Cy3 by
the gold surfaces. Substrate (a) produced a peak signal o f 15,000 counts (PMT
voltage = 400V). The substrate (b) yielded the signal o f 12,000 counts. This
suggests that the substrate (a) has PA surface density in the glass gap region 25%
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
higher than the substrate (b). The corresponding pulse shifts were Axp = -12ps for
substrate (a), vs. Axp = -9.5ps for substrate (b). This suggested pulse transit time shift
was 26% larger for substrate (a) than (b). The difference in transit time o f slide (a)
and slide (b) correlated well with the density of the protein on measured with
fluorescence.
16000
14000
12000
w
C
3O
O
^
10000
8000
ignal strip (go ld) re gio n
6000
4000
0
100
200
300
400
500
slide width ( u m )
Figure3. 10 U Vfluorescence profile from the secondary antibody C y3-labelled anti-R-IgG
‘sandwich assay ’ on top o f PA an d R-IgG.
3.3.2.4
Sensitivity Estimation and Affinity Constant Calculation
The pulse shift response vs. the surface density o f IgGs can be conservatively
estimated on the assumption that the CTL surface is covered with a dense R-IgG
monolayer. The R-IgG Fc chain binds with the PA sublayer specifically, hence the
lens shape IgG molecule is likely to be oriented with the long axis up to the surface.
The mass density with a full coverage o f the lens shaped antibody surface has been
estimated to be 6~7ng/mm2 [6, 7] corresponding to around 12ps pulse shift from the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
CTL probe. Since the probe is capable o f resolving 0.2ps shifts, this indicates that the
•y
sensitivity o f the measurements is estimated to be at least lOOpg/mm .
From the binding curve o f R-IgG (0.33nM~ 670nM) to the immobilized protein
A shown in Figure 3.9, we calculated the affinity constant
(K
d
= 1/K).
A
linearization o f the Langmuir isotherm model [8] was used.
A + B
^
= * 4_ =
kd
I B]
AB
(3.3)
I AB]
[ A] [ B]
(3.4)
kd
1
,
[A]
If we assume that the relative surface coverage r = [ 5 ]/[S]max= A r ;, / a t pmzx, where
[B] is R-IgG, the slope o f ([R-IgG]/A r p) vs. [R-IgG], shown in Fig. 3.11, is
1/ a i
ai, and the intercept is 1/ ( at
K). The model gives A rpmax =-11.84ps, and
K d = 4.7 lx l O'9 M. The K d of the binding o f PA with human IgG has been reported
to be 1.62x1 O'9 M [9] and ~10'8 M [10]. Therefore, our result is within the range of
the published results.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
60
50
</>
Q_
40
S l o p e B = 1 /A t
m
©
E
A t max = * 1 1
30
>.
[R- I gG] / Td vs . [R- I gG]
l i n e a r r e g r e s s i o n : Y= A + B X
A*0.39812
B*0.08445
R «0.99927
5
c
d
c
8
0o>
20
-8
-6
I n t e r c e p t A = -1 /(K A tp mtK)
K '1=4 .71 x 1 0 ‘*M
0
100
200
300
400
500
R a b b i t IgG c o n c e n t r a t i o n ( n a n o m o l a r )
600
700
Figure 3.11 (R-IgG concentration/time shift) vs. (R-IgG concentration) from which the
overall affinity constant o f PA an d R-IgG are determined. The affinity constant is estim ated
to be K d = 4 .7 l x l O'9 M.
3.4 Conclusions
We demonstrated that the CTL probe configuration was very sensitive to changes
in permittivity caused by BMIs at a liquid/solid interface. The permittivity change
was due to the displacement of water by forming a protein complex layer on the
transmission line interface. To directly measure the changes in permittivity caused
by BMIs was difficult. We avoided this problem by measuring the fast pulse transit
time shift and the pulse shape change caused by the permittivity change. The pulse
transit time shift and pulse shape change were quantified using the cross-correlation.
The probe operation demonstrated that the CTL technique could not only detect if
the binding interactions took place at the interface (Biotin-antibiotin, R-IgG-anti RIgG), but also perform a real time measurement of the interaction. The binding
isotherm o f PA and R-IgG was investigated in real time using a wide range o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
sample concentrations (from 0.3nM~700nM). The probe sensitivity was estimated to
y
be ~100pg/mm for the R-IgG binding to PA experiment. The affinity constant was
estimated to be K d = 4.71xl0'9 M, well within the published range.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
3.5 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
T. Dubrovsky, A. Tronin, C. Nicolini, "Determination o f orientation o f the
IgG molecules in immoblized Langmuir monolayers by means o f binding
with fragment specific anti-immunoglobulin antibodies," Thin Solid Films,
vol. 257, pp. 130-133, 1995.
B. Lu, M. R. Smith, R. O'Kennedy, "Immunological activities o f IgG
antibody on pre-coated Fc receptor surfaces," Analytica ChimicaActa, vol.
331, pp. 97-102, 1996.
C. H. Collie, J.B. Hasted, D.M. Ritson, "The dielectric properties o f water
and heavy water," Proceedings o f the Physical Society, vol. 60, pp. 145-160,
1948.
G. T. Hermanson, Bioconjugate Techniques. San Diego: Academic Press,
1995.
T. Dubrovsky, A. Tronin, S. Dubrovskaya, S. Vakula, C. Nicolini,
"Immunological activity o f IgG Langmuir films oriented by protein A
sublayer," Sensors and Actuators B, vol. 23, pp. 1-7, 1995.
A. T. T. Dubrovsky, S. Dubrovskaya, S. Vakula, C. Nicolini, "Immunological
activity of IgG Langmuir films oriented by protein A sublayer," Sensors and
Actuators B, vol. 23, pp. 1-7, 1995.
"Cole-Parmer Technical Library, Principles in adsorption to polystyrene
(Bulletin N o .6 ,2nd edition)."
S. Ross, J.P. Olivier, On Physical Adsorption. New York: Interscience, 1964.
http.7/www.affinity-sensors.com/. "Kinetics Analysis Protein A and Human
IgGl Interaction."
A. Mallia, P. Smith, Immobilized Affinity Ligand Techniques'. Wiley, 1992.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
Chapter 4 Dielectric Coplanar Transmission Line Probe Modeling
and Sensitivity Study
4.1 Introduction
In the previous chapter, the feasibility studies demonstrated the probe
detected the biomolecular interactions (BMIs) at the transmission line solid/liquid
interface. We hypothesized the interfacial dielectric change was mainly due to the
displacement o f water molecules close to the liquid/solid interface.
In this chapter, we present an electromagnetic model and provide an
experimental verification o f the CTL interface sensing mechanism. Because the
antigen and antibody binding forms a monolayer o f protein complex [1], we
simulated the BMI binding events with a formation o f a thin non-dispersive
dielectric layer at the CTL interface. We verified our electromagnetic model by the
deposition o f polyelectrolyte multilayers onto the CTL. The electromagnetic
behavior of the CTL was modeled by solving Maxwell’s equations for all the field
components with a two-dimensional finite difference frequency domain (FDFD)
numerical full-wave solution [2, 3], We experimentally verified the electromagnetic
model with the experiments of depositing monolayers in a controlled manner onto
the CTL probe surface. The theoretical model and experimental verification are
consistent and explains why the CTL probe is sensitive to the permittivity changes at
the liquid/solid interface.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
4.2 Theoretical Model for CTL Probe
The pulse propagation in a transmission line is modeled by a linear system
[4]. The output pulse is
f ( t , z) = ^
| F(co, z = 0)H (c o Y jo>,d(o
(4.1)
where F(ca, z = 0) is the Fourier transform o f the input pulse / (t, 0), The transfer
function H(co) is
H (o ) = Exp[-k0(a(co) + jfi(co))z]
(4.2)
k0 = cdI c where co is the radian frequency ,c is the speed o f light in vacuum, a(co)
is the attenuation factor, J3(a>) is the phase propagation factor. This transfer function
captures the dispersion characteristics o f the transmission line (i.e. the frequency
dependence o f the propagation characteristics).
The dispersion depends on geometry o f the structure (i.e. the waveguide
dispersion) and the frequency dependence of the dielectric property o f different
regions (i.e. the material dispersion). Let the unperturbed waveguide transfer
function be H 0(a ) = exp[-k0(a 0(co) + jf50(co))z\ , and the perturbed waveguide
transfer function b e H x(co) = exp[-k0(a x(co) + jP x(co))z] .The probe senses changes in
the propagation A/(co) = Aa(co) + iAfi((o) = a ^(m )~ a 0(co) + i/?,(co) - if3o(co) (4.3)
The output pulse shape will be distorted if either A a or Af5, or both, are dispersive.
If A y is not dispersive the pulse shape stays the same, then
F(co,z = 0)jf/,(<y)« F(co,z = 0)H {(co) = F(co,z = 0)H 0((o)Exp[-koAyz]
(4.4)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
and therefore, f ( t , z ) » e~k°Aazf 0(t - x, z ) with x = A fiz / c . Thus, the pulse is simply
delayed by a factor o f x , which is proportional to the change in the propagation
constant.
If A y is dispersive, the case is more complicated. The 2cm long aqueous
region on top o f the CTL introduces significant dispersive factors into the system.
The dispersion o f the CTL must be calculated by solving Maxwell’s equations for all
the field components, including the longitudinal propagation components in the
CTL. The quasi-static solutions are not applicable for this case, because it implicitly
assumes the dominant mode has field components only transverse to the direction o f
propagation.
Therefore, we used a two-dimensional finite difference frequency domain
(FDFD) numerical full-wave solution technique [2, 3]. This numerical full-wave
solution can handle an arbitrary cross section o f the waveguide. The contour path
feature models extremely thin layers while retaining a grid spacing that is orders o f
magnitude larger than the thickness o f the layers [2, 3]. When the water is absent in
the sample region, all the materials around the CTL are lossless. In such a case, the
quasi-static solution o f the effective dielectric constant was consistent with the
FDFD model described below with < 2% variation.
Figure 4.1 shows the CTL probe geometry used for the electromagnetic
model. The geometric and dielectric properties o f each layer (substrate, conductors,
polyelectrolyte layers and liquid) were discretized with a staggered finite difference
numerical scheme. The CTL was modeled by placing a magnetic symmetry
boundary wall at the symmetry plane. H alf o f the signal strip was 200pm and the gap
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
region was 50 pm. The CTL glass substrate was modeled as a 400jam thick region
with Sr=4.6. The metallic electrodes on top o f the glass were assumed to be infinitely
thin perfect conductors. A 100pm thick sample region was placed above the
electrodes covering both the gap and electrodes. The sample region was filled with
water. Perfect electric conductors were placed at the remaining three boundaries. It
was confirmed that they were far enough from the gap regions such that the
propagation characteristics o f the CPW mode could be calculated with sufficient
accuracy. The computation window was chosen to be 800pm x 375pm, which was
confirmed to be a sufficient size for assuming the electromagnetic field to be zero
outside of the window. The numerical grid spacing was 12.3 pm x 1.6pm, resulting in
a 32 x 501 grid size.
Initially, the 100pm thick sample region was filled with water. The
dispersion properties for water was assumed to follow the Debye model at 25°C [5,
6]. Debye model was discussed in the Appendix A. The remaining region on top o f
the sample region was filled with poly methyl-methacrylate (PMMA) (er= 2.7) [7].
The dielectric perturbation from sample region was modeled by a thin planar nondispersive dielectric layer (er= 2.5 or 5) build-up in the sample region. The dielectric
layer build-up started from the CTL electrode surface. The dielectric layer thickness
increased from 0 to 100pm displacing water in the sample region.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
PMft/IA
Non-dispersivedielectric laver
g=50um
s=400um
Glass
Figure 4.1 CTL p ro b e geom etry f o r the electrom agnetic modeling
The dispersion properties have been calculated for the probe geometry as a
function o f thickness o f a layer with s r = 2.5 and e r = 5. The complex propagation
constant is y = k 0 ^js eff . The real and imaginary parts o f the effective dielectric
constant s eff are shown as a function of frequency and the layer thickness for
s r = 2.5 in Figure 4.2. Figure 4.2 confirms that the effective dielectric constant
(e eff) o f the dispersion o f CTL is mainly due to water. As the non-dispersive
dielectric layer displaces the water, the e eff becomes less dispersive. Thin layers
close the CTL surface change the dispersion significantly. The model clearly shows
why the probe is intrinsically sensitive to the small dielectric changes close to the
CTL liquid/solid interface. The s r = 5 case is similar to s r = 2.5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
1420 GHz
12-
0 nm
15 GHz
100 nm
10-
10 GHz
500 nm
w
I
6-
5 GHz
5 urn
425 urn
75 nm
1 GHz
100 nm
R eal ( e j
Figure 4.2 Real and imaginary parts o f the effective dielectric c o n s t a n t s = y as a function o f
frequency and the layer thickness fo r Sr = 2 . 5 . The effective dielectric constant was calculatedfrom
the FDFD model.
The phase velocity p is calculated from the FDFD model, and plotted as a
function o f dielectric constant and thickness o f the thin dielectric layer in Figure 4.3.
The thin layer with lower dielectric constant lowers the effective dielectric constant
o f the transmission line, so the phase velocity increases as the layer thickness
increasing. The slope A/? / AI o f the curve is large when the layer thickness is
changed close to the surface. As the layer growing thicker, the slope is decreased,
indicating the probe becomes less sensitive to the layer thickness change. This again
confirms that the CTL probe was intrinsically sensitive to the dielectric changes at
the interface.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
0.55 0.50 -
o
^
0.45 -
;>
0.40 -
oo
jo
3
Q_
0.35 0.30 0.25 0.20
0
20
40
60
80
100
T h i c k n e s s / *im
Figure 4.3 N orm alized pu lse ph ase velocity as a function o f dielectric constant an d thickness o f the
thin dielectric layer.
The transfer function o f the CTL was then obtained from the FDFD model,
and the propagation o f a pulse could be easily simulated using the Fast Fourier
Transform. We used a transform limited baseband input Gaussian pulse with a
32GHz full-width half-maximum (FWHM) o f bandwidth corresponding to duration
o f 13ps FWHM. In addition to the transfer function o f the transmission line, the
bandwidth of the measurement system was set at 32GHz with a single pole lowpass
filter to account for bandwidth limits of the measurement instrument, probes and
cables.
This Gaussian pulse propagated through three sections o f the CTL probe;
first, a 1.7cm long CTL section with air on top, then a 2cm long CTL sample section
with water on top, and last 1.7cm long CTL with air on top. The Gaussian pulse
after propagating through the CTL probe overlapped with the measured real pulse
output (Figure 4.4). The FFT propagation model was a good approximation to the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
experiment. The differences are mainly due to the pulse used in experiment was not a
Gaussian, and the reflections in transmission line were not taken into account in the
model.
0.02
0.00
- 0.02 -
0)
3o>
-0.04 -
§
-0.06 -
3
-0.08 -
a.
- 0 .1 0 -
- 0 .1 2 -
G aussian Pulse(Thoery)
Input P u l s e ( M e a s u r e d )
-0.14 200
400
600
800
1000
Time/ ps
Figure 4.4 Com parison o f the theoretical m odel an d m easured pu lses through CTL p ro b e with w ater
in the flu id chamber. The theoretical m odel is from sim ulating a J3ps FW HM Gaussian pu lse
propagatin g through the CTL with dispersion properties calculated from the FDFD m odel
The dispersive effects o f the CTL on the pulse propagation were calculated as
a function o f pulse bandwidth and attenuation vs. propagation distance. The results
are shown in Figure 4.5 and Figure 4.6. Note that for a 2cm long propagation length,
which was the length o f the sample channel, the bandwidth o f the input pulse was
reduced from 32GHz to around 5GHz due to the large attenuation o f water at high
frequencies. The factor exp[-k0ocx(a>)z] in the transmission transfer function
attenuated the high frequency components, which limited the bandwidth o f the pulse
and led to significant pulse broadening for the input pulse. The result suggested that
we could use a longer input pulse (with the FWHM>30ps) for the experiment. The
attenuation factor also brought a 15dB signal attenuation to the system. For an input
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
pulse on the order o f 1 volt, a 15dB attenuation in pulse could be easily resolved with
an oscilloscope that had a sensitivity o f lmV.
35
30
N
X
25
o
£
20
■cO
(0
CO
75 nm
50 jam
25 nm
5 nm
0 nm
o .o o o
0.005
0.010
0.015
0.020
Length / m
Figure 4.5 Change in the bandwidth as a function o f propagation length an d thickness o f the thin
dielectric layer with the relative perm ittivity o f er=2.5.
mo
■
in
H
i
o
*—
25 nm
_j
-1 0
5 nm
- 500nm
= 100nm
0 nm
•16
0 .0 0 0
0.005
0 .0 1 0
0.015
0.020
Propagation Length/ m
Figure 4.6 Pulse attenuation as a function ofpropagation length an d the thickness o f the thin
dielectric layer with the relative perm ittivity o f er=2.5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
The change in the output pulse transit time and shape was calculated from the
cross-correlation. The probe sensitivity was estimated from these results. The pulse
time shift ( A r ) and cross correlation ratio ( T ) are shown as a function o f dielectric
constant and thickness of thin dielectric layer in Figure 4.7(a) and (b). The pulse
time shift ( A r ) scales linearly with the pulse propagation length. Whereas the
change o f pulse shape T shows little change between a propagation length o f 1cm
and 2cm. We are most interested in detecting monolayers that are about h = 10 wm
thick, because it is approximately the thickness o f the protein conjugate complex [1],
A r has and approximately linear dependence in the lOOnm film thickness range, i.e.
A
t
/
Lh « 0.041/w /(nm.cm) fo rs r =2.5 . For propagation lengthL = 2cm and film
thickness h = 10 nm , then A r = 0.8 p s . The feasibility test in Chapter 3 showed that
the time domain transmission (TDT) measurement system time resolution ((3 a) was
at least 0.2ps (Section 3.3.2.3). In summary, our TDT measurement set-up is able to
resolve the dispersion changes from the interfacial permittivity change within lOnm.
100
-
£ r= 2 . 5
80 -
1
60-
V
<
40 -
20
-
0
20
40
60
80
100
Thickness /
Figure 4 .7(a) A
t
as a function o f dielectric constant an d thickness o f thin dielectric layer.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
1.00
0.95 -
0.90 -
0.85 -
0.80 -
0.75 -
0.70
0
20
40
60
80
100
T h i c k n e s s / urn
Figure 4 .7(b) T as a function o f dielectric constant an d thickness o f thin dielectric la yer.
4.3 Experimental Verification of the Model-Layer By Layer Polyelectrolyte
Deposition
To verify our theoretical modeling o f the CTL probe, we performed a well
controlled polymer layer deposition experiment. In the early 90s, Gero Decher
showed that it was possible to deposit layers of oppositely charged polymers onto the
substrates with the molecular-level control [8, 9]. The alternative deposition o f
anionic polymer and cationic polymers on the charged surface leads to the formation
o f a film called ‘polyelectrolyte multilayer’. During each layer deposition step, the
polyelectrolytes from the solution are electrostatically attracted by the oppositely
charged polyelectrolytes that are deposited on the substrate in the previous step. The
thickness per layer is typically between 1 and lOnm depending on several
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
parameters, including the ionic strength of the polyelectrolyte solution, the pH, and
the molecular weight o f the polyelectrolyte.
The well studied anionic poly (sodium 4-styrenesulfonate) (PSS) and cationic
polyallylamine hydrochloride (PAH) monolayers were chosen for our experiments
[10,11]. Studies have shown the film thickness o f the PAH/PSS pair increases
linearly with the number o f deposited layer pairs. Cationic polyethylenimine (PEI:
MW= 60,000 Sigma Cat. 181978), PSS (MW= 70,000 Sigma Cat. 243051), and
PAH (MW= 70,000 Sigma Cat. 283223) were obtained from the supplier and used
without further purification. Polyelectrolyte solutions were prepared in ultrapure
deionized water. The layer thickness are controlled by the polyion concentrations
and the salt content o f the deposition solution [12, 13]. Two polyion deposition
experiments were performed. One experiment used polyelectrolyte solutions with a
concentration of 0.25% in 0.1M NaCl solution and the other used 0.5%
polyelectrolyte in 0.03M NaCl solution.
Time domain measurements were performed with the measurement system as
described in Chapter 3. The transmitted pulse data was recorded from the reference
channel and sample channel. The pulse data was processed to calculate for the cross
correlation ratio T and the pulse transit time shift A r . Before each polyelectrolyte
deposition sequence, the baseline drift was measured with both channels exposed to
phosphate buffer saline (PBS pH=7.4 + 0.1, Fisherbiotech, BP399-500, 40 times
diluted) for 1 hour. The baseline was calculated from the cross­
correlation f s (t) \1>RS ® f re/(t) |/>bs. This baseline drift was negligible with the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
variation (a) less than 0.02ps, which means the system was able to resolve changes
in the transit time o f 0.06ps (3a).
The same CTL probe fixture was used in this polyelectrolyte experiment as in
the feasibility test (Chapter 3). The glass in the gap regions o f the probe was made
hydrophilic by exposing the CTL surface to oxygen plasma for 2 minutes to create a
high density o f hydroxyl groups. The PEI was used to form the first adhesion layer
on the glass substrate.
The polyelectrolyte multilayers were built up as follows. A cationic PEI
solution was injected into the sample channel to displace the PBS. By stopping the
flow, the sample channel was exposed to the polyelectrolyte solution for 15 minutes.
Then the surface was rinsed by flowing PBS (O.lml/min) for 15 minutes. The same
exposure-washing cycle was repeated, but alternatively using anionic PSS and
cationic PAH solutions after the first layer PEI deposition.
4.4 Results and Discussions
The cross correlation ratio T showed step-like changes, which correlated with
the CTL exposure to polyelectrolyte, and PBS washing. T was
during the PBS
buffer rinse, and ~ 0.7 during the exposure to polyelectrolyte solutions. This
suggested the dielectric constant o f polyelectrolyte solution and PBS buffer were
different. The difference in permittivity was due to the variation o f salt
concentrations. Refer to the Appendix C. for calculating the permittivity o f a saline
solution. The time-delay A r for each layer was taken after the PBS wash at the
inflection point where the correlation ratio T switched back to 1. The polyelectrolyte
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
deposition results showed a steady increase in time shift A r as the multilayers built
up. The raw data of the two experiments are shown in Figures 4.8 and Figure 4.9.
20
40
60
100
80
0.5
1.05
0.25%, 0.1 M NaCl
0.0
120
PBS
PBS
PBS
PBS
-
1.00
0.95
in
-0.5 -
Q.
0.90
>.
jo
a>
a>
73
0.85
E
i-
0.80
-
2.0
(O
tr
c
o
j3
£
o
o
-
0.75
-2 .5 -
-3.0
PSS
0
20
PAH
40
60
0.70
PAH
PSS
80
PS
100
0.65
120
Time / Minutes
Figure 4.8 The raw experimental data o f dielectric p ro b e response to the p olyelectrolyte m ultilayer
deposition. The CTL p ro b e was exposed to 0.25% PSS (anionic) in 0. lM N a C l an d 0.25% P AH
(cationic) in 0. IM N a C l alternatively. PBS buffer wash was applied after each exposure. D o tted line
with data poin ts corresponds to time shift, an d smooth line corresponds to correlation ratio.
0.5%. 0.03 M NaC
Time / minutes
Figure 4.9 The raw experimental data o f dielectric p ro b e response to the p olyelectrolyte m ultilayer
deposition. The CTL p ro b e was exposed to 0.5% PSS (anionic) in 0.03M N aC l an d 0.5% PAH
(cationic) in 0.03M N aC l alternatively. PBS buffer wash was applied after each exposure. D o tted line
with data poin ts corresponds to time shift, and smooth line corresponds to correlation ratio.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
Using the time shift after the first PEI layer deposition as the reference, the
pulse transit time shift A r = t
- t pei
for the successive PSS and PAH layers is
plotted as a function of layers in Figure 4.10.
-1.4
0.5%
^
-
0.8
-
-
0.6
-
<
0.25%
-0.4 -
-
0 .2
-
0.0
-
0
1
2
3
4
5
Number of Layers after PEI deposition
Figure 4.10 Time shift A t = T — TPEI f o r the successive PSS an d PAH layers (using the tim e shift
after the fir s t P E I layer deposition as the reference) a function o f number o f layers. The straight line
is the linear regression result from the data. C ircled data poin ts correspond to the time shift m easured
from 0.25% P A H /P SS pair in 0.1M NaCl. Square data points correspond to the tim e shift m easured
from 0.5% PAH/PSS p a ir in 0.1M NaCl.
A decrease in the pulse transit time is observed from the polyelectrolyte
deposition. This is consistent with a decrease in effective permittivity o f the CTL
probe. The probe experimental response to the monolayers is almost linear with the
slope o f 0.25 ps/pair of 0.5% PSS/PAH and 0.18 ps/pair of 0.25% PSS/ PAH
respectively. Assuming the permittivity o f the polyelectrolyte s r = 2.5, our FDFD
theoretical model for the probe were used to predict the thickness o f the dielectric
layer. From the FDFD model, a slope of 0.25 ps/layer (0.5% PSS/PAH)
corresponds to 6nm thick PSS/PAH pair, and a slope o f 0.18 ps/layer (0.25%
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
PSS/PAH) corresponds to 4.5nm thick PSS/PAH pair. Decher and co-workers have
estimated that a PSS/PAH layer (0.5% polyionic concentration in 0.1M N a C l) is
approximately 8nm. Their estimation was made using optical measurements and
assuming an optical index in the 1.48 - 1.50 range [8-9]. Precisely measuring the
polyelectrolyte layer thickness is very challenging since the layer thickness depends
on not only the chemical nature o f the polymer pair, but also the concentration o f salt
added to solutions used for deposition and rinsing. Our 6nm experimental result
roughly agreed with Decher’s 8nm result.
4.5 Conclusions
The electromagnetic field analysis and experimental verification with the
polyelectrolyte confirmed that the CTL probe system was sensitive enough to detect
the change near the solid/liquid interface within ~10nm. The CTL is most sensitive
to dielectric changes above the surface in close approximation o f 1OOnm. The CTL
probe is intrinsically sensitive as a signal transducer to investigate the specific
binding or molecular recognition events at or near the interface.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
4.6 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
T. Dubrovsky, A. Tronin, C. Nicolini, "Determination o f orientation o f the
IgG molecules in immoblized Langmuir monolayers by means o f binding
with fragment specific anti-immunoglobulin antibodies," Thin Solid Films,
vol. 257, pp. 130-133,1995.
C. M. Arft, "Finite difference frequency domain method for thin, lossy,
anisotropic layers in waveguides," in Electrical and Computer Engineering.
Davis: University o f California, 2004, pp. 200.
C. M. Arft, A. Knoesen, "An efficient finite difference frequency domain
method including thin layers," Microwave Optical Technique Letters, vol. 43,
p p . 40-44, 2004.
R. E. Collin, Foundations o f Microwave Engineering: McGraw-Hill, 1966.
C. H. Collie, J.B. Hasted, D.M. Ritson, "The dielectric properties o f water
and heavy water," Proceedings o f the Physical Society, vol. 60, pp. 145-160,
1948.
P. Debye, Polar molecules: chemical catalogue company, 1929.
"http://www.dowcoming.com/content/photonic/75-l 021-01 .PDF."
G. Decher, J. Schmitt, "Fine-Tuning o f the ultrathin multilayer films
composed o f consecutively alternating layers o f anionic and cationic
polyelectrolytes," Progress in Colloid Polymer Science, vol. 89, pp. 160-164,
1992.
Y. Lvov, H. Hass, G. Decher, H. Mohwald, A. Kalachev, "Assembly o f
polyelectrolyte molecular films onto plasma-treated glass," Journal o f
Physical Chemistry, vol. 97, pp. 12835-12841, 1993.
E. V. Hubsch, B. Senger, G. Decher, J-C. Voegel, P. Schaaf, "Controlling the
growth regime of polyelectrolyte multilayer films: changing from exponential
to linear growth by adjusting the composition o f polyelectrolyte mixtures,"
Langmuir, vol. 20, pp. 1980-1985, 2004.
J. Ruths, F. Essler, G. Decher, H. Riegler, "Polyelectrolytes I:
polyanion/polycation multilayers at the air/monolayer/water interface as
elements for quantitative polymer adsorption studies and preparation o f
hetero-superlattices on solid surfaces," Langmuir, vol. 16, pp. 8871-8878,
2000 .
J. Ruths, F. Essler, G. Decher, and H. Riegler, "Polyelectrolytes I:
Polyanion/polycation multilayers at the air/monolayer/water interface as
elements for quantitative polymer adsorption studies and preparation o f
hetero-superlattices on solid surfaces," Langmuir, vol. 16, pp. 8871-8878,
2000 .
M. Schonhoff, "Layered polyelectrolyte complexes: physics o f formation and
molecular properties," Journal o f Physics-Condensed Matter, vol. 15, pp.
R1781-R1808,2003.
A. Stogryn, "Equations for calculating the dielectric constant o f saline water,"
IEEE Transaction on Microwave Theory and Techniques, pp. 732-736, 1971.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
[15]
[16]
C. Malmberg, A. Malyott, "Dielectric constant o f water from 0 to 100°C,"
Journal o f Research o f the National Bureau o f Standards, vol. 56, pp. 1-8,
1956.
G. S. Smith, W. R.Scott, "The use o f emulsions to represent dielectric
materials in electromagnetic scale models," IEEE Transaction on Microwave
Theory and Techniques, vol. 38, pp. 323-334, 1990.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
Chapter 5 Characterization of Biomolecular Interactions Using
Coplanar Transmission Line Probe Coupled With Fluorescence
Technique
This chapter focuses on an experimental study designed to test if the CTL
probe could differentiate between the biomolecular interactions (BMIs).
In the
previous chapters, we demonstrated by experiments and theory that the dielectric
CTL probe was sensitive enough to detect the BMIs at a solid/liquid interface. In this
chapter, we couple the CTL with the fluorescence technique, and also atomic force
microscopy (AFM) to derive more information o f BMIs at the molecular level. The
results indicate that the CTL probe could not only detect if a specific binding takes
place but also differentiate between two BMIs.
5.1 Introduction
Dielectric measurements o f the biological solutions exhibit distinctive spectra in
the microwave frequency range [1-5]. This was the motivation for our exploration o f
the CTL sensor to extract molecular information o f the BMIs. In this chapter, we
investigated two model assays that both involved the binding to the immunoglobulin
target molecules. The binding interaction took place at different functional parts o f yimmunoglobulin (IgGs). The protein A(PA) to Rabbit-IgG (R-IgG) interaction
involves the Fragment crystallizable (Fc) portion o f IgG, which is in contrasts to the
biotin to anti-biotin mouse IgG interaction that involves the Fragment antigen
binding (Fab) portion o f the IgG [6]. Refer to Figure 5.1 for details. These two
ligand-protein interactions therefore expected to bring different physical and
chemical properties changes, since they involve the interaction with different
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
characteristics in the protein structure. In our experiment, we used i) PA binding to
R-IgG, and ii) biotin binding to anti-biotin y-immunoglobulin obtained from mouse
(a-biotin IgG).
The differences o f these two BMI assays were assessed with the CTL probe. The
fluorescence technique was used to independently determine the relative density o f
IgG molecules bound to their corresponding ligands on the surface. In both assays,
the target molecules (R-IgG, and a-biotin IgG) were labeled with Cy3 fluorescent
tag at comparable levels. The fluorescent tag was not necessary for the operation o f
the CTL probe. The CTL results, when combined with the fluorescence results,
indicated that the CTL probe was more sensitive to the Fc binding (PA to R-IgG)
than the Fab binding (biotin to anti-biotin). This conclusion was reached because the
results that the ratios of pulse shift to surface density (measured by fluorescence)
were about one order o f magnitude larger for R-IgG binding to PA-modified surface
than a-biotin IgG binding to a biotinylated surface.
r m rr-
9 9 ^^
99
Target Molecule Rabbit IgG Fc portion binds with PA
►Protein A Ligand Molecules
►Target Molecule anti-biotin mouse IgG Fab portion binds with B-BSA
►B-BSA Ligand Molecules
Figure 5.1 Two bioassay binding interactions take p la c e a t different functional p a rts o f yimmunoglobulin (IgGs). The protein A (PA) to Rabbit-IgG (R-IgG) interaction involves the Fragm ent
crystallizable (Fc) portion o f IgG. The biotin to anti,biotin mouse IgG interaction involves the
Fragm ent antigen binding (Fab) portion o f the IgG.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
5.2 Experimental
Two ligand molecules, protein A (PA) or biotinylated bovine serum albumin (BBSA) was immobilized by covalent bonding to the activated glass gap regions
straddling the signal strip on the coplanar probe. In the case o f PA, it was covalently
linked to glass surface by reductive animation [8], The glass gap region surface was
activated with aldehyde trimethoxysilane (AMS) (PSX1050 ‘Bio-Conext’ from
United Chemical Technologies). First, the CTL substrates were subjected to oxygen
plasma cleaning (5 minutes 50W, 40 mTorr PO 2 ) followed by submersion in 1%
AMS in anhydrous methanol solution overnight. Substrates were then rinsed with
copious amounts o f methanol to remove excess AMS, and the surfaces were
annealed in an oven in air at 120 °C for 30 minutes. PA (Pierce 21184) was
immobilized by incubating the AMS surface with ca 100 g/ml PA IX phosphate
buffer saline (PBS) solution (pH 7.4) in the presence o f 0.1 mM NaBH 4 .
In the case o f B-BSA, it was covalently linked to the glass surface using a more
reactive silane: trimethoxysilylpropyl isocyanate (TMSP-NCO). The slides were
immersed in 1% TMSP-NCO in toluene and were reacted overnight. The slides were
rinsed clean with iso-propanol, and the surface was annealed in an oven at 65°C for
10 minutes. B-BSA (0.5mg/ml in 0 .1M carbonate buffer pH=9) was incubated on
the isocyanates surface for 4-8hours.
After the surfaces were functionalized, the CTL substrates were stored till use in
a refrigerator at approximately 8 °C. The CTL was assembled by soldering the
microwave connectors (SubMiniature version A, SMA) to a copper plate. Refer to
Section 3.3.1 in Chapter 3 for details. The same time domain transmission
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
measurement system was used here as described in Chapter 3. The data was taken
every
20
seconds in real time during the experiment.
A typical experimental procedure was as follows. First, the PBS carrier solution
(40 times diluted) was injected into both sample channel and reference channel on
the probe. The pulse transit time shift between the transmitted pulses in the signal
channel f s(t) |ra? and the reference channel f ref (t) |/>^ was calculated from the cross
-correlation f s(t) \PBS <8 >f reJ(t) |/>Bs. This pulse transit time shift served as the baseline.
The variation o f the signal difference (a) was always smaller than 0.02ps, which
means the system could resolve changes in the transit time o f 0.06 ps (3a). A BSAPBS solution at high concentration BSA (lmg/ml) was flowed through the sample
channel for 20-30 minutes before each binding experiment to block non-specific
binding sites on the probe surface and in the fluidic delivery channels. BSA does not
specifically bind to the ligand molecules, but may be physically adsorbed onto the
surfaces o f the probe.
Following this blocking step, the BSA solution was displaced by the PBS carrier
solution. The sample solution with target molecules (R-IgG, or a-biotin) was then
injected into the sample channel, while leaving the reference channel with PBS
carrier solution throughout the experiment. Target molecule R-IgG was anti-goat
rabbit IgG Cy3 conjugate (Sigma C2821, F/P=7.0). Target molecule a-biotin was
anti-biotin mouse IgG Cy3 conjugate (Sigma 5585, F/P=7.3). All sample solutions
were made by using the PBS carrier solution to dilute the original sample to different
concentrations. The sample solution flowed through the CTL at a constant flow rate
o f 0.05ml/min. for 20 minutes. PBS was then flowed over the sensor to remove
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
excess sample solution (O.lml/min. flow rate). The formed PA+R-IgG, and biotin+abiotin molecular complexes had long dissociation rates, and remained intact during
the PBS wash. After the PBS wash, the next sample (with the same target molecule
at a higher concentration) was injected.
From the cross-correlation o f the two pulses transmitted through the sample
channel and the reference channel, the change in pulse transit time A r p was
calculated. Distinctive switch over points were observed from the cross-correlation
ratio T, which corresponded to small changes in the dielectric constant in the bulk
liquid in the sample chamber. The switchover points were used to detect when the
target molecules presented in the sample region and also when the pure PBS rinsed
out the excess target molecules. As the surface reactions took place, the pulse transit
time decreased indicating that the pulse propagated faster. This was consistent with
a decrease in the effective dielectric constant o f the transmission line caused by
BMIs.
5.3 Results
The probe’s in-situ response to non-specific binding is shown in Figure 5.2. The
measurement was taken when the probe surface was exposed to a high concentration
o f BSA in PBS (1 mg/ml). This experiment clearly showed the sensor response to
bulk dielectric changes in the sample channel. The change in the transit time o f the
probe to the BSA bulk solution was Axp = -1.5 ps. The BSA was then washed out o f
sample channel, and the change o f the transit time was Atp = -0.1 ps. The BSA does
not bind with the ligand molecules on the CTL surface. Most o f the BSA was rinsed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
away and the probe showed low level o f sensitivity to the remaining o f non-specific
protein. As stated early, this essential step blocked the non-specific binding sites on
the CTL surface and in the fluidic delivery tubes.
0.2
0.0
-
0.2
-0.4
-
0.6
-
0.8
-
1.2
-1.4
-
1.6
-
1.8
9
12
15
18
21
24
Time/minutes
Figure5.2 CTL response to bovine serum albumin (BSA). Block solution (Img/ml) BSA was
introduced to the CTL bio-reaction channelf o r 10 minutes fo llo w e d b y a PBS buffer wash. The
reference channel w as f d le d with PBS buffer during the experiment. The CTL transient response to
BSA bulk solution w as A r = -1.5ps. The pu lse shift after PBS wash w as Ar = -0. lps.
The PA modified CTL to detect the specific binding with R-IgG was tested first.
Sample solutions o f R-IgG with increasing concentrations were injected into the
chamber. Between each sample, the unreacted R-IgG molecules were washed out o f
the chamber with PBS. The antibody solutions with increasing concentrations were
introduced consecutively until the surface reached saturation. The saturation was
reached when no further change in the pulse transit time was measured. Two
surfaces were tested; substrate (a) with a low and substrate (b) with a high density o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
AMS-PA molecules on the surface. For substrate (a), after the exposure to a 5 pg/ml
R-IgG, the CTL surface reached saturation. The total pulse transit time shift was
-1.lps (Figure 5.3(a)). For substrate (b), the saturation point was reached at 50pg/ml.
The total pulse transit time shift was -12ps (Figure 5.4(a)). A higher concentration o f
R-IgG to reach the saturation suggested a higher density o f protein complex formed
on the CTL surface. The higher density o f PA+R-IgG complex resulted in a larger
change in the effective permittivity o f the CTL.
The fluorescence measurement result suggested the pulse shift was scaled with
concentration o f ligand molecules (AMS-PA). Cyanine (Cy3) fluorescence labels
were used to verify independently the PA to R-IgG binding and also to get a relative
calibration o f the PA densities o f the two surfaces. The R-IgG was Cy3 labeled and
the fluorescence label to protein ratio (F/P) ratio was 7. Fluorescence intensity
profiles o f Cy3 were measured with a fluorescence imager (Axon Instruments
GenePix 4000A). The fluorescence distribution profiles at a cross section o f the
substrate (a) and substrate (b) are shown in Figures 5.3(b) and 5.4(b), respectively. In
each figure, two peaks with high fluorescence intensity confirmed that PA+R-IgG
binding took place predominantly inside the glass gap regions o f the probe. The
substrate (a) yielded fluorescence o f 4,500 counts (PMT voltage = 500V), and for
substrate (b) o f 15,000 (PMT voltage = 400V; 70,000 counts when scaled to 500V).
The two measurements were not performed with the same PMT voltage because the
detector for substrate (b) saturated at 500V. The fluorescence result from substrate (b)
was scaled to 70,000 counts at PMT=500V from the experimentally determined
scaling factors. The fluorescence result on the substrate (b) was estimated to be at
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
least 15 times higher than substrate (a), which correlated well with the
11
times
difference in transit time shift measured from the CTL probe.
50
60
70
80
1 .0 5
1.00
j/m l, At= \ 0 .5 3 i
$
—►
S)
- 0 .8 -
0 .9 5
; L>
.
(1ug/m l, AT=-0.898ps)
( 5 u £ /m l A T = -1 .1 9 1 p s)
v
/
W i
0 .9 0
0
3cn
V)
1
8“ i
0
0
)
I
i
o'
*
0 .8 5
0 .8 0
50
60
70
80
tim e/m inutes
Figure 5.3(a) The CTL p ro b e surface m odified with Protein A ligand(substrate(a)) w as exposed to
Rabbit IgG (R-IgG) as the analyte. The CTL response was continuously m onitored during the flo w o f
R-IgG (Sigma C282I). The pu lse shift reading (in the p lo t) w as taken after the buffer wash an d before
the next sample.
5000
4500 4000 -
W
C
3
3500 3000 -
q
2500 -
^
2000
-
1500 -
1000
-
5 00 -
100
150
200
25 0
300
35 0
40 0
slide width(um)
Figure 5.3(b) Fluorescence profile fro m C y3-labelled R-IgG (Sigma C2821) on PA m odified
substrate (a).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
0
10
20
30
40
50
60
70
.005
R-IgG (50ng/ml)
At = -4 .4 8 p S
,
oa>
w
8
Q.
Va
Q.
'P'
<
R-IgG (250ng/ml)
=-6.85ps \
-8
R-IgG (50ug/ml)
=-10.68ps
«
a)
D
CL
0 .9 9 0
Ax
Ax
*
-10
R-IgG (2.5ug/m
=-8.99ps
-12
0.9 8 5
Ax
-14
-16
0 .9 8 0
0
10
20
30
40
50
60
70
time/minutes
Figure 5.4(a) The CTL p ro b e surface m odified with Protein A ligand(substrate (b)) w as exposed to
Rabbit IgG (R-IgG) as the analyte. The CTL response w as continuously m onitored during the flo w o f
R-IgG (Sigma C2821). The pu lse shift reading (in the plot) w as taken after the buffer wash an d before
the next sample.
16000
glass regions
14000
12000
c
3
<3
>
z>
10000
8000
signal strip (gold) region
6000
4000
0
100
200
300
400
500
slide width (um)
Figure 5.4(b) F lu o re sc e n c e p r o f ile fr o m the C y 3 -la b e lle d -R -Ig G (S ig m a C 2 8 2 1 ) on to p o f P A
m o d ifie d su b s tr a te (b).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
The BSA-biotinylated surface was tested with a Cy3-labeled a-biotin IgG from
mouse (Sigma 5585) with a similar experimental protocol as for the PA to R-IgG.
The a-biotin mouse IgG with increasing concentrations (500ng/ml, 1pg/ml, 2.5pg/ml,
and lOpg/ml) was introduced till the surface reached saturation, refer to Figure 5.5(a).
The total time shift from a-biotin binding onto the biotinylated probe surface Atp
was -1. lps. The fluorescence measurement indicated a peak signal o f 28,000 counts
(PMT voltage o f 500V, Figure 5.5(b)). The F/P ratio o f the a-biotin IgG was 7.3.
0 .0
-
a -b io tin
(5 0 0 n g /m l, (A t)= -0 .2 5 9 p s )
- 0 .2 -
§
O
A
; \
'■
.0 4 -
■
8
I
1
I ;
1
a-b io tin
( lu g /m l, (At)= -0 .4 3 5 p s )
1
\ f
«
"
I
k \\
,3
V
1’
V
<i -°-6 '
w
\ </ '
M
u* *■
J
a-biotin
/ \
(2 -5ug/m l, (A t)= -0 .7 7 2 p s )
\
a-b io tin
\
I
(5 u g /m l, (Ai)= - 0 .8 1 3 p s )
a-b io tin
\ (1 O ug/m l, (At)= - 1 .0 4 p s )
u
- 1 .0 -
— I-------- 1-------- 1-------- 1-------- 1-------- 1-------- 1-------- 1-------- 1-------- 1-------- 1-------- 1---------- 1------ 1—
30
40
50
60
70
80
90
I
100
tim e /m in u te s
Figure 5.5(a) The CTL p ro b e surface m odified with biotinylated bovine serum albumin (B-BSA). The
ligan d w as exposed to anti-biotin mouse y- immunoglobulin IgG (a-biotin) as the analyte. The CTL
response w as continuously m onitored during the flo w o f a-biotin (Sigma 5585) solution o f increasing
concentrations. The pu lse shift reading (in the plot) was taken after the buffer wash an d before the
next sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
30000
BSA-biotin s u rfa c e
25 0 0 0 -
20 0 0 0 -
8
15000-
10000 -
5000-
50
100
150
200
250
300
350
400
450
s lid e w id th (u m )
FigureS. 5(b) The U Vfluorescence profile o f the cy3 labeled a-biotin mouse IgG on the B-BSA
surface w as obtain ed using Axon Instruments GenePix 4000A. The two fea tu re p ea k s with high
fluorescence intensity confirmed that biotin binding was predom inantly inside the glass gap regions
on the CTL. The substrate yieldedfluorescence o f 28,000 counts (PM T voltage = 500V).
Table 5.1 Summary o f Fluorescence and CTL Probe Results
Bioassay Under Test
PA+R-IgG (substrate (a))
Fluorescence Result
(Fluorescence Counts)
4,500
CTL Probe Result
(Pulse Transit Time Shift )
-l.lps
PA+R-IgG (substrate (b))
70,000
- 1 2 ps
B-BSA+anti-biotin mouse IgG
28,000
-l.lps
5.4 Discussions
The experimental results demonstrated that the CTL probe could differentiate
between two different BMI events. The two BMI results could be compared, because
two binding antibodies under test (a-biotin mouse IgG and rabbit-IgG) had similar
molecular weights, binding affinities with the ligands, and the F/P ratios o f the two
target antibodies (R-IgG and a-biotin mouse IgG) were close. The density o f proteins
was estimated from florescence counts (28,000 for BSA-biotin surface, 4,500 for PA
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
surface). The fluorescence results indicated the PA+R-IgG binding assay had lower
surface protein density. The CTL measurement results showed both binding
reactions caused the pulse transit time shift o f -l.lp s.
Our results indicated that the interfacial dielectric property was perturbed more in
PA+RIgG system than in the biotin+a-biotin system. We speculate that it is because
o f the strong hydrophobic interactions created in PA+R-IgG (Fc portion) interaction
which lead to a larger protein conformation change and a more effective
displacement of water molecules. The dielectric change sensed by the CTL probe
was the collective effect o f all factors caused changes in the atomic or molecular
dipole moments. The factors associated with the dipole moments are charge
distribution, state o f protonation and polarity o f the protein molecules, and the
dynamics o f water molecules.
According to the molecular model o f PA, the fragment B o f PA forms two
contacts with the Fc portion o f the antibody IgG. The major contact involves
residues from both helices of fragment B o f PA, and the residues from the CH 2 and
CH 3 domains o f antibody Fc portion. The interaction is predominantly hydrophobic
[8-13]. This hydrophobic interaction forms the PA+R-IgG conjugate complex, and
causes the major changes o f the protein conformations and the dynamics o f ‘bound’
water molecules. The second contact o f PA to R-IgG involves a sulfate ion and is
mainly a polar interaction. The polar interaction also causes changes o f the dipole
moments o f the protein molecules.
The biotin to a-biotin interaction involves Fab portion of the a-biotin IgG
antibody binding with biotin. The Y shaped antibody with two identical binding sites
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
that are complementary to a portion o f the surface o f the biotin molecule. A detailed
examination o f antigen-binding sites o f antibodies shows that they are formed from
several loops o f polypeptide chain that protrude from the ends o f a pair o f closely
juxtaposed protein domains [14-16]. The interactions are mainly polar interactions.
Our CTL results suggested the hydrophobic interaction in PA+R-IgG binding
caused large changes in protein shape, charge distribution, and the dielectric
relaxation frequency o f ‘bound’ water. Hydrophobic interactions refer to the
tendency o f the hydrocarbon groups in proteins to gather and form intermolecular
binding. The hydrocarbons in proteins form intermolecular aggregates in an aqueous
medium in hydrophobic interactions, which causes large protein conformation
change. The ‘bound’ water molecules refer to water molecules forming hydrogen
bonding with the proteins. Owing to the nature o f the strong forces between the
bound water and its neighboring macromolecules, the relaxation frequency o f
‘bound’ water should be lower than that o f pure liquid water. During the
hydrophobic interaction, many ‘bound’ water molecules are ‘freed’, and some ‘free’
water molecules form hydrogen bonding at different sites.
Comparing with the CTL technique, a surface plasmon resonance (SPR)
technique can not distinguish the binding events o f PA to R-IgG vs. biotin to a-biotin.
Both the CTL transit probe and SPR sensors detect BMIs by measuring permittivity
changes at a solid-liquid interface. The CTL transit probe detects permittivity
changes at frequencies that extend to microwave range, whereas SPR detect changes
in the optical range. In the optical range, water and protein structures have a similar
refractive index value and both have little dispersion. Water has a refractive index o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
1.35 and proteins typically have a refractive index o f about 1.5 [17]. For this reason,
different BMIs produce similar permittivity changes in the optical regime.
On the other hand, the dielectric properties characterize the atomic and
molecular dipole moment changes in the material in the microwave frequency range
[2, 7], Studies have shown that biological solutions have several dielectric
relaxations in the microwave frequency regime. (3 dispersion is assigned as the
biomolecular dipole moment relaxation, which is below a few hundred MHz.
8
dispersion was observed in the measurements o f DNA [11] and protein (BSA [12])
solutions, and is assigned as the ‘bound’ water molecular relaxation, which falls in a
few GHz. y dispersion is the free water dipole relaxation, which is around 20 GHz.
The CTL detects the effective dielectric constant change during the BMIs, which is a
collective effect o f changes in all those dipole relaxations. Therefore, the result
contains more molecular signature information.
While the CTL transit probe can differentiate between biomolecular events, it is
not as sensitive as SPR probes. The difference is by one or two orders o f magnitude.
SPR has achieved detection levels on the order o f 0.5-1 Opg/mm2 [18]. In our PA to
R-IgG binding experiment, the change in the pulse transit time scaled linearly with
the surface density o f IgGs, which was measured by florescence. The R-IgG Fc
chain binds with the PA specifically, hence the lens shape IgG molecule was likely
to be oriented with the long axis up to the surface. With the assumption that the
surface was covered by a dense R-IgG monolayer, the mass density has been
estimated to be 6~7ng/mm 2 [19, 20]. The saturated surface produced a 12 ps
decrease in transit time. Our measurement system could resolve at least 0.2ps in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
transient time, so the sensitivity o f CTL probe to PA+R-IgG interactions is estimated
atlOOpg/mm . Following a similar argument, the sensitivity o f the CTL probe to
biotin+a-biotin interactions is 600pg/mm2.
The CTL shows different sensitivity to different BMIs, therefore it is not a simple
organic mass sensor like the quartz crystal microbalance (QCM). The CTL detects
the permittivity changes from the BMIs taking place at the solid/liquid interface.
Changes in the permittivity during the BMIs are related not only to the protein
complex surface density change, but also the protein conformation change, the
hydration property of the interactions, and the dynamics o f water molecules.
Comparing with the layer-by-layer polyelectrolyte deposition measurement
results, the change in the pulse transit time from BMIs was much larger. The layerby-layer polyelectrolyte deposition measurement results showed a pulse transit time
change o f 0.25ps per polymer layer (PAH/PSS pair). Each polymer layer was
approximately 6 nm. The pulse transit time shift was much larger for biotin + a-biotin
interactions (l.lp s). Atomic force microscopy (AFM) result showed the protein
complex layer was around
6
to 15nm thick (Refer to Appendix D for details).
A much larger pulse transit time shift from the BMI suggested that the
permittivity changes during the BMI were larger than the polyelectrolyte deposition.
The polyelectrolyte layer is much less dispersive than the protein complex layer. The
protein-ligand binding interaction is not a simple non-dispersive thin layer
displacement o f water at the interface. The size o f the IgG molecules, the protein
binding complex orientation, and protein protonation may all cause change in
permittivity. A larger pulse transit time shift may also suggest that the dynamics o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
water plays an important role in the BMI events. Water has a large permittivity and
is very dispersive in the microwave regime [21]. Studies have shown that the
‘bound’ water that forms hydrogen bonding with the protein molecules has different
relaxation frequency as oppose to the pure water molecules. During the BMIs, some
‘bound’ water may get ‘freed’ by breaking the hydrogen bonding, and some ‘free’
water may form hydrogen bonding with the protein at different sites. Permittivity o f
water at the CTL interface may be greatly changed, which causes a large change in
pulse transit time.
5.5 Conclusions
In summary, a dielectric coplanar transmission line probe was used to detect
two binding assays (protein A and R-IgG, biotinylated-BSA and a-biotin mouse
IgG). The CTL probe could differentiate between these two BMIs. Our results
suggested the Fc portion of R-IgG binding to PA caused more interfacial dielectric
property change than the Fab portion o f a-biotin mouse IgG binding to the BSAbiotin. The sensitivity o f the CTL probe depends on the BMI under test. The protein
A modified CTL surface can reach the sensitivity o f lOOpg/mm for detection o f RIgGs, and the B-BSA modified CTL probe has the sensitivity o f 600pg/mm2 for
detection o f a-biotin mouse IgGs.
At current point, the CTL probe does not have as high sensitivity as the
optical sensors such as SPR (0.5~10pg/mm2). However, the CTL is able to
differentiate between different binding events, because the aqueous biological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
solutions exhibit more spectra information in the microwave regime than in optical
regime.
AFM results showed that the BSA-biotin+ a-biotin mouse IgG protein
complex layer was on the same order o f thickness as one polylelectrolyte polymer
pair (PAH/PSS). The change o f pulse transit time from the BMI (1.1 ps) was larger
than the polymer (0.25ps). The difference suggested the BMI interaction caused
larger and more complicated dielectric property change than a non-dispersive thin
layer displacing water at the interface.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
5.6 References
[1]
[2]
[3]
[4]
[5]
[6 ]
[7]
[8 ]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
C. G. Essex, M. S. Symonds, R. J. Sheppard, E. H. Grant, "Five-component
dielectric dispersion in bovine serum albumin solution," Physics in Medicine
& Biology, vol. 22, pp. 1160-1167, 1977.
E. H. Grant, R.J. Sheppard, G.P. South, Dielectric behavior o f biological
molecules in solution, 1 ed: Oxford University Press, 1978.
S. Bone, "Dielectric studies o f native, unfolded and intermediate forms o f
beta-lactamase," Physics in Medicine & Biology, vol. 39, pp. 1801-1809,
1994.
J. Hefti, A. Pan, A. Kumar, "Sensitive detection method o f dielectric
dispersions in aqueous-based, surface-bound macromolecular structures using
microwave spectroscopy," Applied Physics Letters, vol. 75, pp. 1802-1804,
1999.
G. R. Facer, D.A. Notterman, L.L. Sohn, "Dielectric spectroscopy for
bioanalysis: from 40Hz to 26.5GHz in a microfabricated wave guide,"
Applied Physics Letters, vol. 78, pp. 996-998, 2001.
B. Alberts, Molecular biology o f the cell, 3rd ed. ed. New York: Garland Pub.,
1994.
H. P. Schwan, K. Foster, "RF-Field interactions with biological systems:
electrical properties and biophysical mechanisms," Proceedings o f the IEEE,
vol. 6 8 , pp. 104-113, 1980.
G. Kronvall, K. M. Grey, R. C. Williams, "Protein A reactivity with mouse
immunoglobulins," Journal o f Immunology, vol. 105, pp. 1116-1123, 1970.
J. Sjodahl, "Structural studies on the four repetitive Fc-binding regions in
protein A from staphylococcus aureus," European Journal Biochemistry, vol.
78, pp. 471-490, 1977.
J. Deisenhofer, "Crystallographic refinement and atomic models o f a human
Fc fragment and its complex with fragment B o f protein a from
staphylococcus aureus at 2.9- and 2.8- A resolution," Biochemistry, vol. 20,
pp. 2361-2370,1981.
B. Lu, M. R. Smith, R. O'Kennedy, "Immunological activities o f IgG
antibody on pre-coated Fc receptor surfaces," Analytica Chimica Acta, vol.
331, pp. 97-102, 1996.
J. Yakovleva, R. Davidsson, M. Bengtsson, T. Laurrell, J. Emneus,
"Microfluidic enzyme immunosensors with immobolised protein A and G
using chemiluminescence detection," Biosensors and Bioelectronics, vol. 19,
pp. 21-34, 2003.
L. Yang, M.E. Biswas, P. Chen, "Study o f binding between protein A and
immunogobulin G using a surface tension probe," Biophysical Journal, vol.
84, pp. 509-522, 2003.
C. L. Homick, F. Karush, "Antibody affinity (3)-the role o f multi valence,"
Immunochemistry, vol. 9, pp. 325-340, 1972.
F. A. Saul, L. M. Amzel, R. J. Poljak, "Preliminary refinement and structural
analysis o f the Fab fragment from human immunoglobulin at 2.0A
resolution," Journal o f Biological Chemistry, vol. 253, pp. 585-597, 1978.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
[16]
[17]
[18]
[19]
[20]
[21]
A. Orlova, E. H. Egelman, "Structural dynamics o f F-actin: changes in C
terminus," Journal o f Molecular Biology, vol. 245, pp. 582-597, 1995.
J. Homola, S.S. Yee, D. Myszka, Opitical biosensors .present andfuture:
Elsevier Science, 2002.
E. Kress-Rogers, "Handbooks o f biosensors and electronic noses: medicine,
food and the environment," 2nd ed: CRC press, 1997.
A. Mallia, P. Smith, Immobilized Affinity Ligand Techniques: Wiley, 1992.
T. Dubrovsky, A. Tronin, S. Dubrovskaya, S. Vakula, C. Nicolini,
"Immunological activity o f IgG Langmuir films oriented by protein A
sublayer," Sensors and Actuators B, vol. 23, pp. 1-7, 1995.
C. Malmberg, A. Malyott, "Dielectric constant o f water from 0 to 100°C,"
Journal o f Research o f the National Bureau o f Standards, vol. 56, pp. 1-8,
1956.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
Chapter 6 Nanoporous Thin Film on the CTL Transmission Line
Probe Surface for BMI Detections
6.1 Introduction
In this chapter, we present the detection o f the biotin to anti-biotin, and biotin to
avidin interactions using a CTL probe with a nanoporous thin film integrated to the
surface. The performance o f the nanoporous film CTL probe was compared with the
probe without the nanoporous film. The nanoporous film CTL probe showed a
selective response to two target molecules, anti-biotin and avidin which are o f
different molecular sizes and both bind with the biotin ligand molecule. This result
suggested it was possible to physically separate the target molecules based on their
sizes.
The porous film is generated by the sacrificial porogen approach developed by
the IBM research center (Almaden, San Jose) [1,2]. This approach utilizes phase
separation o f two component system where one component (e.g., organosilicates
matrix) crosslinks into a network effectively limiting domain growth and controlling
the topology o f the porogen phase (an organic, labile polymeric component) that is
ultimately expelled from the film by thermal decomposition [2, 3]. Some key
material properties that characterize the porous film are pore density, pore size, size
distribution and pore structure [4]. An open structure mesoporous film is used in this
investigation. In an open porous structure, the pores are interconnected and the liquid
can go inside the pores. In this way, a larger surface area is created for the ligand
loading than a plain substrate. The sample fluids containing the target molecules can
be force-flowed through the mesoporous support, so that the target/ligand molecular
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
interaction volume is also increased. The mesoporous support refers to the material
with the average pore size of 2 to 50nm. The geometry o f the porous material could
be adjusted to form a well-controlled uniform pores with different sizes.
For the microwave CTL probe, the signal transduction interface and the BMI
interface do not need to be as tightly coupled as the surface plasmon resonance
(SPR). CTL probe allows layers o f mesoporous material or microstructured surfaces
such as microfabricated columns or beads in the sensing region which is not possible
for SPR. The wavelength o f a microwave signal is on the order o f cm, therefore, the
microstructured material with less than
1 /1 0
o f the signal wavelength will not cause
the wave scattering.
6.2 Experimental
A nanoporous support was fabricated on the CTL probe glass gap surface as
follows. A poly (methyl silsesquioxane) (PMSSQ, GR650F from TechneGlas) was
used as the matrix and a block copolymer of ethylene oxide and propylene oxide
(P I23, Pluronics, BASF) was used as the porogen. l-methoxy-2-propanol acetate
(PMAc) solution (1 wt%) o f PMSSQ/porogen mixtures containing 80 wt% o f PI 23
was coated on glass surface between the CTL gold electrodes using a 75 pm
capillary. In this way, only the glass surface between the electrodes is coated with
the porous material. The samples were fully cured by heating to 450°C at 5°C/min
and hold at 450°C for 2 hours under argon atmosphere. A UV/ozone stripper
(SAMCO, UV-300H), which has a low pressure Hg lamp and an ozone generator,
was used to hydrophilize the surface. After UV/ozone treatment at 100°C for 40min,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
the sample surfaces show static water contact angle (contact angle system, AST,
2500 XE) o f less than 10 degrees. Independent small angle X-ray scattering
measurements showed the samples contain a random, open pore structures of
approximately 30nm diameters [1, 3]. The pore density was 50%. Figure 6.1 shows
the detailed schemes about the porous film formation and the porous film image
obtained using scanning electron microscopy (SEM).
Poly(methyl silsesquioxane) (PMSSQ)
Porogen Mixture
^ p h a s e separation by thermal crosslinking
Nanohybrid
^Therm olysis of porogen phase
Nanoporous Film
SEM Image of nanoporous PM SSQ film
Figure 6.1 Schem atic o f the form ation o f PMSSQ nanoporous film and SEM image o f the porou s film .
Biotinylated bovine serum albumin (B-BSA, Pierce 21346) was immobilized in
the nanoporous support in the following manner. The substrate was first plasma
treated for 5 min at 400 mTorr O 2 100W. The substrate was then submerged in a
slide holder filled a 1% solution o f N-[3 (trimethoxysilyl) propyl]ethylenediamine
(Aldrich 104884 or Gelest #SIA0591) in anhydrous methanol for 12 hours at room
temp. The DSC (disuccinimdyl carbonate) method was used to pre-coat the amines
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
on the surface [5]. The surface was activated by reacting for 3 hours with 25mM
DSC di-(N-succinimidiyle)-carbonate (Aldrich 225827, Fw= 256.17) dissolved in
acetonitrile (Aldrich 75058) 40ml andl mM DIPEA (N, N)-diisopropylethylamine,
(Fw= 129.25, Aldrich 387649) as a catalyst. The substrates were washed with
dimethyformamide solution for 30 minutes, followed by rinsing briefly with
methanol and oven-dried the substrates in 65°C. B-BSA solution with the
concentration of 500(ig/ml in MES buffer pH 5.4 was incubated on the activated
surface for 30 minutes, and was rinse-cleaned with BSA solution, and water.
For comparison purpose, another CTL probe with a plain glass surface in the gap
region was modified with B-BSA in the following way. B-BSA was covalently
linked to the glass surface using a more reactive silane: trimethoxysilylpropyl
isocyanate (TMSP-NCO). The slides were immersed in 1% TMSP-NCO in toluene
and were reacted overnight. The slides were rinsed clean with iso-propanol, and the
surface was annealed in an oven at 65°C for 10 minutes. B-BSA (0.5mg/ml in 0.1 M
carbonate buffer pH=9) was incubated on the isocyanates surface for 4-8 hours [6 ].
Two control glass slides, one had the nanoporous film on it, the other was plain
glass were treated with the same protein immobilizations as their reference probes.
After the protein immobilization, the ligand protein loading density was measured
using the Cy3 labeled extravidin (e-avidin, Sigma E4142, F/P= 4.8). lOpg/ml eavidin was used to spot on the control slide B-BSA surface for 20 minutes. The
surface was then rinsed clean with water. Fluorescence intensity profiles o f Cy3
were measured with a fluorescence imager (Axon Instruments GenePix 4000A).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
We tested the biotinylated CTL probe response to two target molecules, anti­
biotin mouse IgG (a-biotin) and e-avidin. The same time domain transmission
measurement set-up was used here as described in Chapter 3. The data was taken
every 20 seconds in real time. The same experimental procedure described in
Chapter 5 was used to measure the baseline drift. The signal drift was negligible with
the variation (a) smaller than 0 .0 2 ps, which means the system could resolve changes
in the transit time o f 0.06ps (3a). A BSA-PBS solution at high concentration
(1 mg/ml) was flowed through the sample channel for 20-30 minutes before the
binding experiment. BSA does not specifically bind to the ligand molecules, but may
be physically adsorbed onto the surfaces o f the probe solution to block non-specific
binding sites on the probe surface and in the fluidic delivery channels.
Following this blocking step, the BSA solution was displaced by the PBS carrier
solution. Solutions o f a-biotin IgG from mouse (Sigma 5585) with increasing
concentrations in PBS were injected into the reaction chamber. Between each
sample, the unreacted a-biotin was washed out o f the chamber with PBS. The
sample solutions with an increasing antibody concentration (500ng/ml, lpg/m l,
2.5pg/ml, 5pg/ml, and lOpg/ml) were introduced consecutively. After exposure to
the a-biotin IgG, the surface was then exposed to e-avidin (Sigma E4142) in the
same manner as the a-biotin IgG with increasing concentrations (lOOng/ml, lpg/m l,
5pg/ml). In this experiment, we did not focus on the surface saturation, but on the
difference in response between the two CTL probes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
6.3 Results
The B-BSA ligand molecule loading density on the control slides plain glass
surface and the nanoporous film surface was measured using the Cy3 labeled eavidin. For the biotinylated nanoporous film substrate, the fluorescence counts were
56,000 with background o f 4,000. For the biotinylation on the plain glass
substrate,the fluorescence was 26,000 with back ground o f 1,500. The fluorescence
results showed that the ligand protein loading was increased by
2 .1
times with the
nanoporous thin film support. We need to point out here that this result did not
indicate that the porous surface provided higher loading yield than the plain surface,
since different surface modification chemistry was applied. The higher ligand protein
density on the nanoporous film surface could be due to either the surface
modification chemistry or the increased surface area from nanoporous film.
The CTL probe with B-BSA deposited onto the plain glass surface showed a total
time shift Axp o f -lp s from a-biotin binding (Refer to Figure 6.2). The following
exposure to e-avidin produced an additional -0.53ps change in the pulse transit time
(See Figure 6.3). This further change in rin d ic a te d that even though a-biotin IgG
stopped binding to the surface, the extra-avidin binding event continued. More
importantly, it confirmed that both a-biotin IgG and avidin selectively bind to a
BSA-biotinylated planar interface.
For the nanoporous thin film support CTL probe, no measured changes in the
pulse transit time occurred when the CTL was exposed to a-biotin IgG (Refer to
Figure 6.4). In contrast, when the CTL was exposed to e-avidin after the a-biotin
IgG, it produced a pulse transit time shift o f -1.47ps (see Figure 6.5). This result
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
appeared to verify that the nanoporous support selectively responded to avidin vs. abiotin IgG.
0 .0 -
a - b io tin
- 0 .2 -
500ng/m l AT=-0.278ps
jC
<0
a - b io tin
a - b io tin
2.5pg/m l A t= -0.759ps 5pg/m l A t=-0.740ps
-
0.8
a - b io tin
lOpg/ml At=-0.961ps
- 1. 0 -
-
1.2
20
30
40
50
60
70
80
90
100
time/minutes
Figure 6.2 The biotinylated CTL plain glass substrate response to a-biotin mouse IgG (Sigma 5585).
The CTL p ro b e responded to a-biotin IgG with increasing concentrations (500ng/ml, Ipg/ml,
2.5pg/ml, 5pg/m l and 10 pg/ml).
- 0.8
- 1 .0 - -
e-av id in
100ng/m l A t= -1 .3 2 1 p s
- 1.2 -
V)
t
e-av id in
1ug/m l A i= -1 .5 3 5 p s
e-av id in
5 u g /m l A t = - '
- 2 .0 -
100
110
120
tim e /m in u te
Figure 6.3 The biotinylated CTL plain glass substrate response to extra-avidin (Sigma E4142) after
exposed to a-biotin IgG (Figure 6.1). A total time shift o f-1 .5 3 p s was observed after introducing the
extra-avidin with increasing concentrations (lOOng/ml, Ipg/ml, and 5pg/ml) to the substrate.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
0.4-
0 .2
-
a -b io tin
a -b io tin
-D
C
00
I
o
o
*>
no •
C
O “0.2
O
O
\
-
0.8
-
-
1.0
*
^
T
f
f
w
*v
uv
•V
—r~
k
/
i *1^
./
\
/
w i
30
/ ja
*1
H
^
- 0 .6 ■
0
a
1 u g /m l A i= - 0 .2 7 7 p s « *
(V \
j
1
£
)
a --bb io tin
Jl
'P ' -0.4 •
w
10ug/m lA T = - 0 . 1 1 1 p s
5 0 0 n g /m l At= - 0 . 1 7 7 p s
H
----
1
60
— I—
50
40
— i—
80
70
t i m e /m in u t e s
Figure 6.4 The biotinylated nanoporous CTL substrate response to a-biotin mouse IgG (Sigma 5585).
The CTL p ro b e h ad alm ost no response to a-biotin IgG with increasing concentrations (500ng/ml,
Ipg/ml, and 10 jug/ml), an d very sm all pu lse time shift w as observed as com pared with the CTL p ro b e
without porou s film (Refer to Figure 6.2) .
o.o-
*co
8a>
-0.5
e -a v id in (1 0 0 n g /m i)
AT=-0.56ps
(A
8
Q.
I
H
!c
<o
J032)
- 1.0
n
\ A y ./
«/w
e -a v id in (1 u g /m l)
At= - 1 .2 3 p s
* *.V
e -a v id in (5 u g /m l)
A t = - 1 .4 7 p s
—V.
-1 .5 -
— I—
130
140
— I—
150
—«---- 1----'—
— i—
160
170
— i—
180
—I
190
tim e /m in u te s
Figure 6.5 The biotinylated nanoporous CTL substrate response to extra-avidin (Sigma E4142) after
exposed to a-biotin IgG. A total time shift o f -l.4 7 p s was observed after introducing the extra-avidin
with increasing concentrations (lOOng/ml, Ipg/ml, and 5pg/ml) to the substrate.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
6.4 Discussions
We expected the nanoporous film modified CTL probe to show a better
sensitivity than the unmodified probe, since the increased protein density should
cause larger dielectric property changes at the interface. A rather unexpected yet
interesting result showed that the nanoporous filmed CTL probe did not respond to
the first target molecule a-biotin IgG, but only responded to the following target
molecule e-avidin (/l^,=-1.47ps). The CTL probe without nanoporous film
responded to both target molecules {A tp= -\.53ps in total), a-biotin IgG caused -lp s
pulse transit time change, and e-avidin caused -0.5ps pulse transit time change, abiotin IgG and e-avidin both bind to the same ligand molecule B-BSA, but e-avidin
molecule is about 3 times smaller in molecular weight than a-biotin IgG [7, 8 ]. The
molecular weight o f a-biotin IgG is 150,000-170,OOODa., while the molecular
weight o f e-avidin is
6 8
,OOODa.
The difference between the nanoporous film modified CTL probe vs. unmodified
probe suggested that it was possible to use a well-controlled nanoporous support
with sufficiently uniform pore sizes to physically separate two target molecules
based on their molecular size. This experiment suggested that the porous structure
blocked selectively the larger molecules.
The selectivity o f a biosensor is commonly achieved by chemically modifying
the substrate with proper ligand molecules. The ligand molecules selectively
“capture” target molecules from the sample. However, it often happens that there are
more than one target molecules competing with each other for binding with the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
ligand molecules. The nanoporous film provides an additional option to physically
improve the CTL probe selectivity to BMIs by adjusting the pore sizes.
The fluorescence results showed the nanoporous film probe surface yielded
higher protein loading density, but it did not demonstrate the higher protein density
was due to the increased surface area from the nanoporous film. Two different
protein immobilization chemistries were applied to biotinylate the surface. The
higher ligand protein loading density o f the porous CTL could be either due to the
increased surface area from the nanoporous film, or due to the surface modification
chemistry. A better designed experiment should be to use the same protein
modification method. Works have been done in comparing the nanoporous film on a
glass substrate without CTL features. Both porous film glass substrate and plain
glass substrate were modified with B-BSA using GOPS (glycidoxyproplytrimethoxysylane, Gelest SIG5840) activated surface [9]. The result showed the
nanoporous substrate had higher protein ligand yield than the glass substrate.
However, we had difficulties in applying the same chemistry to the porous CTL
probe surface [10]. Other surface modification methods should be explored in the
future. Another concern with the nanoporous film is that the background also
increases after the ligand immobilization. In other words, we get more loading but
the specificity o f the surface does not improve. In this regards, a nano or mesopore
surface is not an improvement unless one can design a fluidics system capable o f
driving the flow through the pores.
The CTL probe with planar silicon surface makes it very suitable to integrate
with nanoporous structures. The signal transduction interfaces and the binding
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
interfaces o f the CTL sensor are not as tightly coupled as in the SPR sensor. The
pore size o f the mesoporous film is on the order o f nanometers, which is much less
than the wavelength o f operation in the microwave regime. The porous region will
not cause scattering, unlike the situation with SPR optical sensors operating at
wavelengths comparable in scale to the pores. This fact also implies CTL probe can
study the binding events in a mixed media, for instance, a gel or cell suspensions.
6.5 Conclusions
This experiment demonstrated that the nanoprous film supported CTL probe
selectively responded to two competing target molecules with different sizes. This
selectivity suggested that it was possible to use a well-controlled nanoporous support
with sufficiently uniform pore sizes to physically separate different target molecules
that compete to bind with the same ligand molecules. Further more, the fluorescence
results showed the application o f nanoporous film to the CTL probe surface might
increase the ligand molecule loading density. This will help to increase the
sensitivity o f the probe.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
6.6 References
[1]
[2]
[3]
[4]
[5]
[6 ]
[7]
[8 ]
[9]
[10]
J. L. Hedrick, R. Miller, C.J. Hawker, K. R. Carter, W. Volksen, D. Yoon, M.
Trollsas, "Templating nanoporosity in thin-film dielectric insulators,"
Advanced Materials, \ ol. 10, pp. 1049-1053, 1998.
A. Knoesen, G. Song, W. Volksen, E. Huang, T. Magbitang, L. Sundberg, J.
Hedrick, C. Hawker, R. Miller, "Porous organosilicates low-dielectric films
for high frequency devices," Journal o f Electronic Materials, vol. 33, pp.
135-140, 2004.
C. Lei, Y. Shin, J. Liu, E.J. Ackerman, "Entrapping enzyme in a
functionalized nanoporous support," Journal o f the American Chemical
Society, vol. 38, pp. 11242-11243, 2002.
P. S. Ho, J. Leu, W.W. Lee, Low dielectric constant materials fo r IC
applications, vol. 9: Springer, 2003.
B. J. Hong, V. Sunkara, J. W. Park, "DNA microarrays on nanoscalecontrolled surface," Nucleic Acids Research, vol. 33, pp. 106-114, 2005.
G. T. Hermanson, Bioconjugate Techniques. San Diego: Academic Press,
1995.
A. Mallia, P. Smith, Immobilized Affinity Ligand Techniques'. Wiley, 1992.
B. Alberts, Molecular biology o f the cell, 3rd ed. ed. New York: Garland Pub.,
1994.
J. McMurdie, Q. Chen, S. Angelos, D. Roitman, A. Knoesen, R. Miller,
"Binding surface optimization for a microwave-based biosensor," Poster
presentation at Center o f Polymer Interfaces and Macromolecular Assemblies
Annual Conference 2003.
D. Roitman, "IBM surface modification," Q. Chen, Knoesen, A., Ed., 2005,
pp. Personal Communication.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
Chapter 7 Conclusions and Future Work
7.1 Conclusions
We developed a label free dielectric probe technique to detect the
biomolecular interactions (BMI) at a liquid/solid interface in vivo. The dielectric
BMI probe is based on a microwave coplanar transmission line (CTL) configuration.
Studies performed in other research groups have demonstrated that planar
transmission line structures could measure the bulk dielectric property o f biological
solutions [1, 2]. In this project, we designed a CTL probe to detect specific BMIs at
the transmission line solid/liquid interface.
Comparing with the optical detection technique surface plasmon resonance
(SPR), the new dielectric BMI sensor detects the permittivity change from BMIs in
the microwave frequency regime. The dielectric property of the biological aqueous
solutions have distinctive spectrum features in the microwave regime [3-7].
Therefore, the microwave dielectric BMI probe may be able to provide molecular
structural information as well as to detect the binding events.
Our dielectric CTL sensor was fabricated on a glass substrate with two
ground electrodes adjacent and parallel to the signal electrode. The glass gap region
between the ground and signal electrode was precoated so that ligand molecules
were immobilized on the activated glass surface via covalent bond. The target
molecules in a fluid sample bound with the ligand molecules and formed the protein
conjugate complex at the CTL solid/liquid interface. The BMIs were detected by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
measuring the relative change o f effective dielectric constant ( s
) o f the CTL. The
BMI occurring at the transmission line interface would lower the s eff, since the
aqueous bio-solutions were better dielectrics (lower permittivity) than water alone at
the microwave frequency regime [ 1 , 2 ].
It is important to point out that throughout this project, we did not try to
extract the exact dielectric constant o f the biological samples, but rather focused on
measuring the relative changes in s ^ caused by the BMIs. First o f all, it does not
require the absolute dielectric constant measurement to detect a BMI. Besides, it is
very difficult to measure the absolute value o f dielectric constant. A complicated
modeling of a composite CTL would be needed to extract the absolution dielectric
constant from the measured s eff. Also, the dielectric property o f aqueous solutions is
very temperature sensitive, which means the dielectric property changes with even a
slight temperature variation (< 1°C). This makes it very difficult to take an absolute
measurement.
A time domain transmission (TDT) measurement was chosen in this project
over the frequency domain measurement for two reasons. First, was to avoid the
impedance mismatch termination problem. An impedance mismatch problem will
plague the frequency domain measurement, whereas in the TDT measurement, the
reflections from the impedance mismatch can be eliminated by time-windowing.
Second, the CTL probe structure was designed that BMI occurs at a section o f the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
transmission line. The frequency domain measurement requires a complicated de­
embedding procedure to define the test plane at the BMI section on the CTL. In
contrary, it does not need any calibrations to perform a relative measurement using a
TDT system. The dielectric dispersion changes from BMIs were measured by
comparing the pulse transit time change. Cross-correlation o f a transient sample
pulse vs. a reference pulse was calculated to quantify the changes in pulse transit
time and pulse shape.
Two series of feasibility tests were performed to demonstrate the CTL probe
applications. The probe operation demonstrated that the CTL technique could not
only detect if the BMIs took place at the interface (Biotin with antibiotin, Rabbit-IgG
with anti Rabbit-IgG), but also could perform a real time measurement o f the
interaction (protein A (PA) to Rabbit IgG (R-IgG)). The binding isotherm o f PA and
R-IgG was investigated in real time using a wide range o f sample concentrations
(from 0.3nM~700nM). The probe sensitivity was estimated to be ~100pg/mm 2 for
the R-IgG binding to PA experiment. The affinity constant was estimated to be Kd =
4.71x10'9 M, which was in the expected range [8 , 9].
The high sensitivity in the CTL probe was achieved because the BMI took
place in the gap region between the signal and ground strip where the electric flux
density was concentrated. We hypothesized the interfacial dielectric change was
mainly due to the displacement o f water molecules close to the liquid/solid interface
at a monolayer scale. Water has a large permittivity and is very dispersive in the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
microwave frequency range (80 at low frequencies, and drops to 20 at around 30GHz)
[10,11]. Any effective displacement o f water at the CTL surface will lower the
effective dielectric constant o f the transmission line.
We simulated the BMI binding events by forming a thin non-dispersive
dielectric layer at the CTL interface. An electromagnetic model was presented and
was verified by the polyelectrolyte multilayer deposition experiment. The
electromagnetic model indicated that CTL probe was most sensitive to dielectric
changes above the surface within approximately 1OOnm. For example, a dielectric
thin layer with the thickness of /i= 1 0 nm and dielectric constant o fs r = 2.5,
displacing water at the CTL solid/liquid interface, caused a pulse transit time
shift A t = 0.8ps. Our TDT system with the time resolution o f 0.2ps was able to
resolve such permittivity change within 2nm. This is on the order o f a protein
monolayer. The CTL probe was intrinsically sensitive as a signal transducer to
investigate the specific binding or molecular recognition events at or near the
interface.
Two designed antigen-antibody binding assays were tested in order to test if
the CTL probe could differentiate them. Different functional groups o f the antibodies
were involved in the binding interactions. First was the protein A (PA) binding to yimmunoglobulin obtained from rabbit (R-IgG) at the Fragment crystallizable (Fc)
portion o f the IgG [12]. Second was the biotin binding to anti-biotin y-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
immunoglobulin obtained from mouse (a-biotin IgG) at the Fragment antigen
binding (Fab) portion o f the IgG [13].
The CTL results coupled with fluorescence results demonstrated that the CTL
probe was able to differentiate these two binding events. One BMI (PA to R-IgG)
showed larger transient time shift than the other (biotin to a-biotin IgG). It indicated
the CTL probe technique could provide more BMI information than the other
commonly used label-free technique such as SPR, since SPR could not differentiate
these two binding assays. At present, the probe still can’t reach the sensitivity o f SPR.
The CTL showed different sensitivity to different BMIs, 100pg/mm 2 for PA to R-IgG
binding and 600pg/mm for biotin to a-biotin binding. The sensitivity o f SPR is
around 0.5~10pg/mm2, which is about 2 orders o f magnitude better than CTL
technique.
We also demonstrated it was important to cooperate with different sensing
techniques and develop a complementary set o f diagnostics that could gather more
complete information about a BMI. The pulse transit time shift from the CTL probe
measurement is determined by the protein layer coverage and the dielectric property
o f the protein layer. From single CTL measurements, we could not draw a
conclusion about two different BMIs. Only when coupled with the fluorescence or
SPR technique to get the protein mass coverage information, we could make the
comparison o f the CTL results, and draw more information about the BMIs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
Therefore, the CTL technique provides a complementary detection tool for BMI
study.
The surface imaging atomic force microscopy (AFM) study o f the
biotinylated (BS A-biotin) CTL reacting with extravidin and anti-biotin mouse IgG
showed the thickness o f the protein complex layer after BMI was around 6~10nm.
The pulse transit time shift from the binding interaction was 1. lps. Our
polyelectrolyte deposition experiment showed a 0.25ps time shift per polymer layer
(PAH/PSS pair). And the thickness o f each polymer layer (PAH/PSS pair) was
around 6 nm. Comparing the thickness profile and the CTL response, we found that
the BMIs at the interface brought larger dielectric property change than the
polyelectrolyte deposition. This large dielectric property change from BMIs may be
due to the protein conformation change, and dynamics o f ‘bound’ water and ‘free’
water.
In chapter 6 , nanoporous thin film was applied to the CTL surface. We
compared the response o f a CTL probe with nanoporous poly (methyl silsesquioxane)
film on the glass surface with an equivalent probe with plain glass surface. Both
probes were biotinylated using biotin-bovine serum albumin (B-BSA). The probes
were first exposed to a-biotin IgG and then to extra-avidin. The a-biotin IgG
molecule is approximately three times larger than extra-avidin in molecular weight.
The probe with the glass surface responded to both, but the probe with nanoporous
film only responded to the extra-avidin. The porous film structure blocked the larger
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
a-biotin IgG molecules. This result suggested a possible option to improve the probe
selectivity by physically separate two target molecules, which competitively bind to
the same ligand molecule, based on their physical sizes.
7.2 Future Work
The sensitivity of the CTL probe needs to be improved. One way to improve
the probe sensitivity is to increase the surface loading o f ligand molecules.
Nanoporous thin film as the CTL surface support should be further explored in this
direction. More interesting experiments should be tuning the CTL probe selectivity
by choosing different pore sizes and testing different bioassay interactions with
different size molecules involved.
In the future, the IgG molecules from the same species should be used for the
bioassay interaction comparison. For example, the PA to rabbit IgG and biotin to
anti-biotin IgG comparison, anti-biotin rabbit IgG should be used so the two
bioassays are more biologically similar to make a comparison. Besides, other factors
influencing the binding such as affinity constant, molecular sizes should be tested
systematically. For instance, the PA modified CTL probe could be exposed to
different IgG molecules (human IgG, goat IgG, etc.) with different affinity constants.
The CTL response comparison could give us more information about the interactions
between the PA and the Fc portion of different IgGs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
For this research project, we coupled the CTL technique with fluorescence to
obtain the protein surface density information and estimate the probe sensitivity. In
the future, it would be more accurate if the fluorescence measurement could be taken
right after the reaction since fluorescence intensity decreases with time. Surface
plasmon resonance (SPR) technique has been commonly used as a label-free
technique to study the BMIs. It would be interesting to couple these two label-free
techniques (CTL probe and SPR). The same bioassay such as PA+R-IgG and biotin+
anti-biotin IgG should be tested using SPR. The real-time measurement results and
the affinity constants estimation from SPR should be compared with the results from
CTL technique.
The electromagnetic model o f the CTL probe successfully simulated the
polyelectrolyte layer by layer deposition, and should be further applied to the BMI
systems. The BMI system could be modeled either with a two-layer model, i.e. a
protein layer and a ‘bound’ water layer, or a one-layer model with an averaged
dielectric constant to describe the protein layer with water on top o f the CTL. For the
two-layer model, the ‘bound’ water layer is not well defined. The ‘bound’ water
permittivity could be estimated using the 5-relaxation frequencies observed by other
groups [4]. The thickness of the ‘bound’ water layer should be changed
systematically to observe its influence to the pulse transit time. The thickness profile
o f the protein complex from the AFM measurement should be used for the BMI
system modeling. For the single layer model, an averaged dielectric constant could
be calculated using some models developed to estimate the permittivity o f a mixture.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
One possible approach would be assuming the dielectric constant o f a mixture is
proportional to the volume of each component [14]. The BMI system modeling
should be adjusted based on the real experimental results. Using the BMI system
model, comparisons o f the effective dielectric constant variations from
polyelectrolyte, biotin+anti-biotin, and PA+R-IgG should be studied. The goal is to
find out the relation between pulse transient time vs. film thickness and film
permittivity in the BMI reaction.
The extraction of dielectric constant from the CTL probe is very challenging.
For detection BMIs, it is not necessary to extract the dielectric constant from the
measurement. But studies on the dielectric property o f a BMI system will need to
retrieve the dielectric spectrum. For instance, in the study o f the ‘bound’ water and
‘free’ water kinetics during a protein-protein interaction, dielectric spectra could
provide the relaxation frequency shift information about the 8 -dispersion and 0 dispersion. Also, the comparison of BMI experiment result with the polyelectrolyte
deposition result indicated that BMIs caused larger dielectric changes than the
polymer layer deposition. We are in the opinion that it could be from the dielectric
property change o f water. A possible approach to test this hypothesis would be to
retrieve the dielectric spectra from the experiments, and compare the spectra features
from the polyelectrolyte deposition experiment and the BMI reactions.
Furthermore, as stated in the conclusion that the CTL probe technique has to
couple with fluorescence or SPR technique to make a comparison o f the BMIs. A
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
further development o f the probe technique to make this technique self complete
may be taking frequency domain measurement. The protein coverage information
may be retrieved from a frequency range where the permittivity o f the protein is
constant and proportional to the protein/water content. While a comparison o f the
protein relaxation frequency shift may give the structural information about the
BMIs.
It is more direct to extract the spectra from frequency domain measurement
than from the time domain. However, a calibration scheme needs to be made for
CTL probe in frequency domain measurement, so that the test plane is defined at the
sample section on the CTL after the calibration. The most likely approach would be
using the TRL (through, reflection, and load) calibration [15-17]. The TRL
calibration requires transmission lines with different lengths, and the reflection
(either open or short). Depending on the frequency range of a measurement will
cover, transmission lines with difference length will be needed for TRL calibration.
Last but not the least, an important advantage o f the CTL probe configuration is
that it could be easily integrated with microfluidics and microstructures. The signal
transduction plane and the BMI interaction surface do not need to be tightly coupled.
The wavelength o f a microwave signal is on the order o f cm, therefore, the
microstructured material with less than
1 /1 0
o f the signal wavelength will not cause
the wave scattering. CTL probe allows layers o f mesoporous material or
microstructured surfaces such as microfabricated columns or affinity beads
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
integrated in the sensing region which is not possible for SPR. It is a good future
project for a label-free electronic sensor chip development based on CTL probe and
integrated with the chromatographic separation affinity columns for protein
separation. The protein separation column could be microfaricated in the CTL gap
region or sample channel. The chip could be integrated with microfluidics for
multiplex detections.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ill
7.3 References
[1]
[2]
[3]
[4]
[5]
[6 ]
[7]
[8 ]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
J. Hefti, A. Pan, A. Kumar, "Sensitive detection method o f dielectric
dispersions in aqueous-based, surface-bound macromolecular structures using
microwave spectroscopy," Applied Physics Letters, vol. 75, pp. 1802-1804,
1999.
G. R. Facer, D.A. Notterman, L.L. Sohn, "Dielectric spectroscopy for
bioanalysis: from 40Hz to 26.5 GHz in a microfabricated wave guide,"
Applied Physics Letters, vol. 78, pp. 996-998, 2001.
C. G. Essex, M. S. Symonds, R. J. Sheppard, E. H. Grant, "Five-component
dielectric dispersion in bovine serum albumin solution," Physics in Medicine
& Biology, vol. 22, pp. 1160-1167, 1977.
E. H. Grant, R.J. Sheppard, G.P. South, Dielectric behavior o f biological
molecules in solution, 1 ed: Oxford University Press, 1978.
H. P. Schwan, K. Foster, "RF-Field interactions with biological systems:
electrical properties and biophysical mechanisms," Proceedings o f the IEEE,
vol. 6 8 , pp. 104-113,1980.
S. Mashimo, N. Miura, N.Shinyashiki, T. Ota, "Dielectric study on
molecular motions o f poly(glutamic acid) in aqueous solutions over a
frequency range o f 10e5~10el0 Hz," Macromolecules, vol. 26, pp. 68596863,1993.
S. Bone, "Dielectric studies o f native, unfolded and intermediate forms o f
beta-lactamase," Physics in Medicine & Biology, vol. 39, pp. 1801-1809,
1994.
http://www.affinitv-sensors.com/. "Kinetics Analysis Protein A and Human
IgGl Interaction."
A. Mallia, P. Smith, Immobilized Affinity Ligand Techniques: Wiley, 1992.
C. Malmberg, A. Malyott, "Dielectric constant o f water from 0 to 100°C,"
Journal o f Research o f the National Bureau o f Standards, vol. 56, pp. 1-8,
1956.
C. H. Collie, J.B. Hasted, D.M. Ritson, "The dielectric properties o f water
and heavy water," Proceedings o f the Physical Society, vol. 60, pp. 145-160,
1948.
T. Dubrovsky, A. Tronin, C. Nicolini, "Determination o f orientation o f the
IgG molecules in immoblized Langmuir monolayers by means o f binding
with fragment specific anti-immunoglobulin antibodies," Thin Solid Films,
vol. 257, pp. 130-133, 1995.
B. Alberts, Molecular biology o f the cell, 3rd ed. ed. New York: Garland Pub.,
1994.
B. M. Kolundzija, Electromagnetic modeling o f composite metallic and
dielectric structures: Boston : Artech House, 2002.
S. Wartenberg, "RF coplanar probe basics," Microwave Journal, vol. 46, pp.
20-28, 2003.
R. Marks, R.R. Pantoja, M.J. Howes, J.R. Richardson, R.D.Pollard,
"Comments on 'Improved calibration and measurement o f the scattering
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
[17]
parameters of microwave integrated circuits' [and reply]," IEEE Transactions
on Microwave Theory and Techniques, vol. 38, pp. 453-454,1990.
J. S. Kasten, M.B. Steer, R. Pomerleau, "Enhanced through-reflect-line
characterization o f two-port measuring systems using free-space capacitance
calculation," IEEE Transactions on Microwave Theory and Techniques, vol.
38, pp. 215-217, 1990.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
APPENDIX A Debye Model and Cole-Cole Model
The Debye model was derived to represent the dielectric relaxation o f polar
liquids [1]. Water in liquid states comes closest to exhibiting Debye relaxation.
When a direct voltage applied to a dielectric is suddenly removed, the decay o f
polarization to zero value is not instantaneous but takes a finite time. Similarly the
build up o f polarization takes a finite time interval before the polarization attains its
maximum value. This phenomenon is described by the general term dielectric
relaxation.
In Debye model, it is the collisions between the molecules which inhibit
polarization reorientation when a steady field on a polar liquid is suddenly removed.
Assume that the rate o f decay o f the polarization is proportional to the instantaneous
value o f the polarization, we have
at
= —— giving P = P0 exp(-/ / r)
r
(A. 1)
For an alternating sinusoidal electric field E = E0 exp(jcor)
Now consider a dielectric system in which the total polarization consists o f two types
o f polarizations, the orientational polarization Pu(/), and electronic polarization, Pe
which is assumed to attain its final value instantaneously because the time required
for it to attain the saturation value is in the optical frequency range.
PT(t) = Pu(t) + Pe
(A.2)
Pe = s 0{sx ~ \)E
(A.3)
where s x is the dielectric constant at infinity frequency.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
114
PT = D - s 0E = e0{s r - l ) E
The total polarization is also
(A.4)
The final value attained by the total polarization is
Pt = £o(£ s ~1)E
es
(A. 5)
is the permittivity at low frequencies, where D = e0s sE
From P = D - s aE and D = s 0s rE
P«= £0(£s-Z'>o)E
(A.6 )
Where alternating e-field is E = Emmejal
Then we get
dP{t) _= -l [r„
, „ - s„J E e ja‘ -P {t)}
s 0{es
dt
(A.7)
t
zL
(F
P it) = C e T + s 0 ^
1+
\v >
x-J- m -
(A. 8 )
jC O T
where C is a constant. If time t is sufficiently large when compared with x, the first
term on the right side o f equation becomes so small that it can be neglected and we
get the solution for P(t)
P(t) = s 0
1+
(A.9)
jC O T
rewrite it in terms o f Pu and Pe
m
m
= s ,( e . -
=l
+ £ ‘,( e ■ 7 °
1+
1+
e
JCOT
E .‘ m
(A . 10)
JCOT
.
-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CA.11)
115
The instantaneous value of flux density D = £os*Emejml
From Eqn. (A.4) and Eqn.(A.l 1), we get
( g ,~ g j
1+
(A. 12)
jW T
(A. 13)
£ = £«, +
(A. 14)
Where e0 is the free space permittivity = 8.854* 10' 12 F/m and £>is the relative
permittivity or dielectric constant. Equations (A. 13) and (A. 14) are known as Debye
equations and they describe the behavior o f polar dielectrics at various frequencies.
Cole and Cole showed that, in a material exhibiting Debye relaxation a plot o f e ”
against e each point corresponding to a particular frequency yields a semi-circle [2 ].
Refer to Figure A .I. This can be easily demonstrated by rearranging the Debye
equations:
(A. 15)
£
(g, +Q
Figure A. 1 Cole-Cole diagram displaying a sem i-circle fo r Debye equations fo r e*
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
116
Theories o f other types o f dispersion phenomena also lead to equations o f the
same type form. The original Debye theory o f dispersion and absorption was
developed for polar gases and diluted solutions o f polar liquids. However, there is a
considerable amount o f experimental evidence to indicate that Debye model does not
describe the observed frequency dependence correctly in some cases. The observed
behavior generally consists in a considerably broader frequency range o f dispersion
and absorption together with a smaller maximum value of s '" than is predicted from
Debye model.
An alternative method for displaying dielectric data is the Cole-Cole equation,
or complex permittivity plot [3]. For systems with a single relaxation, the plot takes
the form o f a semicircle, the center o f which lies on the abscissa. For systems with
multiple uniformly distributed relaxations the center o f the semicircle lies below the
abscissa.
e*=£ +
(A. 16)
(1 +jan)r
The Debye model can be presented as a RLC equivalent circuit. Refer to Figure A.2.
The Cole-Cole model is presented with additional term to account for any
conductance paths within the dielectric.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
£' -
(a) Debye model equivalent circuit
(b) Cole-Cole model equivalent circuit
Figure. A. 2 Equivalent Circuit fo r D ebye M odel and Cole-Cole M odel
Grant & Bateman 4 s work indicated that the dielectric dispersion o f urea and
acetamide in aqueous solution can be reproduced fairly well by either two Debye
relaxations or one Cole-Cole relaxation. There is no preference with one over the
other [4].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
118
References
[1]
P. Debye, Polar molecules: chemical catalogue company, 1929.
[2]
K. Cole, R. Cole, "Dispersion and absorption in dielectrics," Journal o f
Chemical Physics, vol. 9, pp. 341-351,1941.
[3]
R. Cole, "Evaluation o f dielectric behavior by time domain spectroscopy
(complex permittivity)," Journal o f Physical Chemistry, vol. 79, pp. 14691474, 1975.
[4]
J. B. Bateman, G. F. Evans, P.R. Brown ,C. Gabriel, E.H. Grant, "Dielectric
properties o f the system bovine albumin : urea : betaine in aqueous solution,"
Physics in Medicine & Biology, vol. 37, pp. 175-182, 1992.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
119
APPENDIX B Maxwell-Wagner Interfacial Polarization
Consider a heterogeneous system consisting o f two or more materials with
different permittivity and conductivity. Because o f the difference in dielectric
properties, there will be a non-uniform distribution o f free electronic charges across
the interface between the materials. In order to maintain current continuity across the
whole system, a charge carrier discontinuity will give rise to the formation o f charge
double layers at the interface between the dissimilar materials. Such a build-up o f
charge gives rise to an induced frequency-dependant dielectric relaxation. Such a
phenomenon is commonly referred to as a Maxwell-Wagner polarization [1 ,2 ]. The
basic system used for modeling consists o f a two layer heterogeneous system
represented by a parallel plate capacitor with two dielectrics o f different thicknesses,
permittivities and conductivities between the plates.
Figure B. 1 M axw ell-W agner m odel as two capacitors in series and equivalent circuit
To derive the dielectric property (s) vs. frequency (to) due to the interfacial
polarization that exists between two layers o f dielectric materials (ey and ei) with
different conductivity (ci and
0 2 ).
Let dj and d2 be the thickness o f two materials.
When a voltage C is applied across the two dielectrics, the voltage across each
dielectric will be distributed, at t= 0 +,
V10 = V
10
°2
C,+C2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(B .l)
120
V20= V
20
Cl
Cx + C 2
(B.2)
'
when a steady state is reached at t= oo, the voltage across each dielectric will be
V*= V
R,
Rx + i ?2
(B 3)
V- - V ^
The charge stored in each dielectric will change during the transition period. At t=0,
the charge in each layer will be equal:
0 . = e * = C,Vn = C2F21 =
V
(B.4)
at t=oothe charge in each layer will be
O
=C V
1 lM
V
Rx + R2
C P
(B.5)
Qi« c2v2x = 2 2 v
=
2
r x+ r 2
2M
during the transition period the change in the stored charge in each layer is
AQ - Q . - Q , = K C | (^
- ^
>
;)
Aft - a. -a, ■vc2J^xQ+'c,)
The admittance o f the equivalent circuit is
Y
Y\Y2
Yx +Y2
_
0 + 7 ^ i)0 + y^2)
R x + R 2 + jcoRxR2(C, + C2)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(B.6)
<BJ)
We further define r = — +
R,+R,
Y
l - Q ) r,r2 + (o r(r, + r2) - ya>(l - (o r,r2) +
1 + fflV
1
e?
(B.9)
+ r2)
The admittance o f the capacitor with the real dielectric is
Yeq = jo)C0e ' = jcoC0( e '- je " )
,, _
(B .ll)
[(ri + r 2 ) - r ( l - f i ) 2 r ,r 2)]
^
C 0 (i?,+ i?2)
1 + <y2 r 2
1
g' = ----------------- tVl
g"
1
= —
_
[1 L1
o) 2t xt 2 +
coC0(R)+ R2)
1
Ci = ^ £ ° £ i
d,
with
r> _
C,
=
^ £ o£ 2
<y2 r(r, + r 2)]
+ <wV
(B12)
(B.13)
Rx = - ^ A ct,
and
D_
J2
2
A,^—A<j 2
Maxwell-Wagner polarizations often occur at relatively low frequencies in
the region o f l~10M Hz depending on the exact nature o f the interface. Because o f
the heterogeneous nature o f many biological systems, Maxwell-Wagner interfacial
effects can often be significant. Examples o f systems that are likely to exhibit
Maxwell-Wagner dispersions are biological membranes [2, 3], and solutions
containing biological tissues.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
122
References
[1]
[2]
[3]
A. V. Hippel, Dielectric materials and applications. New York: John
Wiley&Sons, 1954.
G. G. Raju, Dielectrics in electric fields. New York: Macel Dekker,Inc, 2003.
E. H. Grant, R.J. Sheppard, G.P. South, Dielectric behavior o f biological
molecules in solution, 1 ed: Oxford University Press, 1978.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
123
A ppendix C Calculation for Permittivity o f Saline Water
The permittivity of a polar liquid is calculated from Debye model provided
the static permittivity s s , relaxation tim e r , and e m are known. The static relative
permittivity ( e s) and relaxation time ( r ) o f a saline solution can be calculated using
the equation derived by Stogryn [1].
s A T , N ) = s,{T fi)a (N )
2 ttt(T, N ) = 2nr(T,0)b(N, T )
(C .l)
Function a(NJ and b(N,T) are given as
3
a(N) = 1.000 - 0.2551N + 5.151 x 10 - 2 N 2 - 6.889 x 10 " 3 N
b(N, T)
(C.2)
= 0.1463 x 1 0 “ 2 N T + 1.000 - 0.04896AT- 0.02967 N 2 + 5.664 x 10 “ 3 N 3 (C.3)
where T is temperature (°C), N is normality o f the saline solution. The
equation works for 0 < T < 40"C,0 < N < 3
£s (T,0) is the static dielectric constant o f water, which can be calculated
using the equation derived by Malmberg and Maryott (1956) [2].
s s(T,0) = 8 7 .7 4 0 -0.4008T + 9.398x 1 0 ’4 T 2 -1 .4 1 0 x 10~6 r
3
(C.4)
for o < r < i o o ° c
Smith and Scott provided a model for calculating the conductivity o f saline
solution at three different ranges o f normality for temperature between 10~25°C [3].
cr(T, N ) = cr(25" C, N)[ 1 + c(N)(T - 25)]
for 15°C < T < 25°Cand 0.1 < A <1.0
where
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(C.5)
124
o-(25°C, N) = N(12.653 - 8.974W1' 2 +11.385A - 10.403W3'2 + 4.7224W2 - 0.82288JV5'2)
(C.6)
and C( N) = 0.0193
References
[1]
[2]
[3]
A. Stogryn, "Equations for calculating the dielectric constant o f saline water,"
IEEE Transaction on Microwave Theory and Techniques, pp. 732-736, 1971.
C. Malmberg, A. Malyott, "Dielectric constant o f water from 0 to 100°C,"
Journal o f Research o f the National Bureau o f Standards, vol. 56, pp. 1-8,
1956.
G. S. Smith, W. R.Scott, "The use o f emulsions to represent dielectric
materials in electromagnetic scale models," IEEE Transaction on Microwave
Theory and Techniques, vol. 38, pp. 323-334, 1990.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
125
APPENDIX D Surface Imaging of Antigen-Antibody Interactions on Biotinylated
CTL Probe Using Atomic Force Microscopy
This section focuses on the surface imaging study of the biotinylated (BSAbiotin) coplanar transmission line (CTL) surface reacting with extravidin (Cy3
labeled) and antibody (anti-biotin mouse IgG Cy3 labeled) using the atomic force
microscopy (AFM). The AFM image results provide us information about the height
o f the bio-conjugate pairs on the transmission line surface as well as more insight
about the topography of the surface.
D .l Experimental
The CTL was fabricated on a glass substrate coated with gold using
photolithography. The BSA-biotin was deposited on the glass gap regions between
the gold signal and ground strip of the CTL using the silaxation. The extravidin-Cy3
conjugate (Sigma.E4142) and anti-biotin mouse IgG-Cy3 conjugate (Sigma. C5585)
with the concentration o f 50pg/ml were incubated on the different CTL channels for
30 minutes in room temperature. Then the surface was carefully rinsed with
phosphate saline buffer (PBS) and dried in air. The AFM measurement was
performed using the molecular force probe 3D controller from Asylum Research. All
images were obtained in a contact mode with the scan speed o f 1Hz. The AFM tip
spring constant was O.IN/m and the tip was positioned at the lOOum wide glass gap
region on the CTL before imaging. The AFM images were taken on three different
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
126
surfaces, a biotin-avidin half-wet surface, a biotin-avidin dried surface and a biotinantibiotin dried surface.
D.2 Results and Discussions
a. BSA-biotin surface with extravidin (half-wet is very different from the
dry)
One o f the interesting features the biotin surface after the bio-conjugate
binding reaction is that the dried surface behaves different from the half-wet
surface. One measurement was taken on a biotin- avidin surface right after
the reaction followed by the PBS buffer rinse. The surface was half wet. The
height o f the protein layer was planned to be measured in the AFM contact
mode with the AFM tip with a high voltage power to ‘dig’ into the protein
layer till reaches the hard surface (glass). The protein layer thickness should
be the ‘digging’ depth o f the AFM tip. This ‘digging’ method only worked
for the half-wet protein surface, and was not successful for the dried biotinavidin, and biotin-antibiotin surfaces. The dried protein surfaces show protein
aggregates with specific ring-shaped features, the heights o f the aggregates
were measured and found to be compared with the height measured using the
‘digging’ method on the half-wet surface.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
127
pm
Figure D. 1 The AF M scan image o f BSA-biotin surface reacted with extravidin-cy3
conjugate (half-wet). N otice the darker square area at the right side corresponds to the A F M
tip scratching mark on the protein surface. The height o f the protein layer is ~6nm.
6
tra c e lin e a In Im ag e (blue)
tra c e lin e b in im a g e (red)
4
2
E
c
£
oo>
0
■2
•4
0
20
40
60
80
100
120
140
160
lateral distance (nm)
Figure D. 2 The line-plots o f the traces denoted in the image o f BSA-biotin and extravidin. The
height o f the BSA-biotin and avidin p a ir is ~6nm.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
128
b. BSA-biotin surface with extravidin (dry)
The AFM tip did not leave ‘scratches’ on the dry surface, but o-ring
shaped protein aggregates were observed instead. The dimension and height
of the aggregates were measured instead.
0.5
1.0
1.5
2.0
2.5
pm
Figure D. 3 The A F M scan image o f BSA-biotin surface rea cted with extravidin-cy3 conjugate (dried)
5
4
3
2
1
•2> o
<D
■1
•2
•3
0
50
100
150
200
lateral distance (nm)
Figure D .4 The line-plots o f the traces denoted in the image o f BSA-biotin and extravidin. The height
o f the BSA-biotin and avidin p a ir is ~6nm which is the sam e as the result from the w et surface.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0.0
0.5
1.0
1.5
2.0
|jm
Figure D. 5 Another scan image o f the BSA-biotin an d extravidin surface. N otice the o-ring
featu res on the surface. The outer diam eter o f ring is ~90nm. N otice the ring sh aped p ro tein
cluster is sm aller f o r avidin since an avidin molecule is sm aller than an antibody.
c. BSA-biotin surface with anti-biotin mouse IgG
The dried biotin-antibody (anti-biotin) surface was investigated, similar
o-ring shaped protein clusters were observed.
4
3
2
1
0
0
1
2
3
4
pm
Figure D .6 AF M height image scan f o r BSA-biotin surface rea cted with anti-biotin mouse IgG - c y 3
conjugate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
130
-
trace a (blue)
trace b (red)
E
c
-10
0
50
100
150
200
250
300
la te ra l d is ta n c e (nm )
Figure D. 7 The line-plots o f the traces denoted in the image o f BSA-biotin an d anti-biotin. The
height o f the BSA-biotin and anti-biotin p a ir is ~17.5nm.
5
1.0
Figure D . 8 A zoom-in AF M scan image f o r the BSA-biotin and anti-biotin surface. N otice the oring sh apedfeatu re on the surface. The outer diam eter o f the ring was m easured to be ~150nm.
The AFM images o f the biotin-anvidin and biotin-antibody surfaces
confirmed that a protein layer was formed on the BSA-biotin surface from
the binding reactions. The antibody layer thickness was around 17.5nm from
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
131
AFM measurement, which is suggesting that the lens-shaped anti-biotin IgG
molecules could be oriented with long axis vertical to the surface. The
antibody IgG is a lens-shaped molecule with the long axis o f 15nm, and short
axis of 3 -4 nm, whereas the BSA molecule is a globular molecule with a
diameter o f 4-5 nm.
The height o f biotin-avidin pair was ~ 6 nm, since avidin is a much smaller
molecule than antibody.
The o-ring feature on the surface are from the protein aggregates,
similar features from different protein were observed in prof. Liu’s group
before. One mechanism o f forming such type o f o-rings could be the globular
BSA molecule with biotins on the surface bind with anti-biotin around it, and
formed the center o f the protein clusters, and more antibodies gathers around
it, and formed the o-ring shape protein aggregates.
Acknowledgement:
Special thanks for Maozi Liu and Prof. Gangyu Liu for from Chemistry Department
at UC Davis for providing assistance in AFM measurements.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
2 839 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа